avx512f

package
v0.0.0-...-3878f85 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Jul 23, 2017 License: MIT Imports: 1 Imported by: 0

Documentation

Overview

THESE PACKAGES ARE FOR DEMONSTRATION PURPOSES ONLY!

THEY DO NOT NOT CONTAIN WORKING INTRINSICS!

See https://github.com/klauspost/intrinsics

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

func AbsEpi64 

func AbsEpi64(a x86.M128i) (dst x86.M128i)

AbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := ABS(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm_abs_epi64'. Requires AVX512F.

func AddRoundSd 

func AddRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

AddRoundSd: Add the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := a[63:0] + b[63:0]
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VADDSD'. Intrinsic: '_mm_add_round_sd'. Requires AVX512F.

func AddRoundSs 

func AddRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

AddRoundSs: Add the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := a[31:0] + b[31:0]
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VADDSS'. Intrinsic: '_mm_add_round_ss'. Requires AVX512F.

func CmpEpi32Mask 

func CmpEpi32Mask(a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

CmpEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmp_epi32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpEpi64Mask 

func CmpEpi64Mask(a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

CmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpEpu32Mask 

func CmpEpu32Mask(a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

CmpEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmp_epu32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpEpu64Mask 

func CmpEpu64Mask(a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

CmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpPdMask 

func CmpPdMask(a x86.M128d, b x86.M128d, imm8 byte) (dst x86.Mmask8)

CmpPdMask: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VCMPPD'. Intrinsic: '_mm_cmp_pd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpPsMask 

func CmpPsMask(a x86.M128, b x86.M128, imm8 byte) (dst x86.Mmask8)

CmpPsMask: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VCMPPS'. Intrinsic: '_mm_cmp_ps_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpRoundSdMask 

func CmpRoundSdMask(a x86.M128d, b x86.M128d, imm8 byte, sae int) (dst x86.Mmask8)

CmpRoundSdMask: Compare the lower double-precision (64-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	k[0] := ( a[63:0] OP b[63:0] ) ? 1 : 0

	k[MAX:1] := 0

Instruction: 'VCMPSD'. Intrinsic: '_mm_cmp_round_sd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpRoundSsMask 

func CmpRoundSsMask(a x86.M128, b x86.M128, imm8 byte, sae int) (dst x86.Mmask8)

CmpRoundSsMask: Compare the lower single-precision (32-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	k[0] := ( a[31:0] OP b[31:0] ) ? 1 : 0

	k[MAX:1] := 0

Instruction: 'VCMPSS'. Intrinsic: '_mm_cmp_round_ss_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpSdMask 

func CmpSdMask(a x86.M128d, b x86.M128d, imm8 byte) (dst x86.Mmask8)

CmpSdMask: Compare the lower double-precision (64-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC

k[0] := ( a[63:0] OP b[63:0] ) ? 1 : 0

k[MAX:1] := 0

Instruction: 'VCMPSD'. Intrinsic: '_mm_cmp_sd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpSsMask 

func CmpSsMask(a x86.M128, b x86.M128, imm8 byte) (dst x86.Mmask8)

CmpSsMask: Compare the lower single-precision (32-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC

k[0] := ( a[31:0] OP b[31:0] ) ? 1 : 0

k[MAX:1] := 0

Instruction: 'VCMPSS'. Intrinsic: '_mm_cmp_ss_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CmpeqEpi32Mask 

func CmpeqEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpeqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmpeq_epi32_mask'. Requires AVX512F.

func CmpeqEpi64Mask 

func CmpeqEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmpeq_epi64_mask'. Requires AVX512F.

func CmpeqEpu32Mask 

func CmpeqEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpeqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmpeq_epu32_mask'. Requires AVX512F.

func CmpeqEpu64Mask 

func CmpeqEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmpeq_epu64_mask'. Requires AVX512F.

func CmpgeEpi32Mask 

func CmpgeEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgeEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmpge_epi32_mask'. Requires AVX512F.

func CmpgeEpi64Mask 

func CmpgeEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmpge_epi64_mask'. Requires AVX512F.

func CmpgeEpu32Mask 

func CmpgeEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgeEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmpge_epu32_mask'. Requires AVX512F.

func CmpgeEpu64Mask 

func CmpgeEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmpge_epu64_mask'. Requires AVX512F.

func CmpgtEpi32Mask 

func CmpgtEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgtEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmpgt_epi32_mask'. Requires AVX512F.

func CmpgtEpi64Mask 

func CmpgtEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmpgt_epi64_mask'. Requires AVX512F.

func CmpgtEpu32Mask 

func CmpgtEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgtEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmpgt_epu32_mask'. Requires AVX512F.

func CmpgtEpu64Mask 

func CmpgtEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmpgt_epu64_mask'. Requires AVX512F.

func CmpleEpi32Mask 

func CmpleEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpleEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmple_epi32_mask'. Requires AVX512F.

func CmpleEpi64Mask 

func CmpleEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmple_epi64_mask'. Requires AVX512F.

func CmpleEpu32Mask 

func CmpleEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpleEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmple_epu32_mask'. Requires AVX512F.

func CmpleEpu64Mask 

func CmpleEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmple_epu64_mask'. Requires AVX512F.

func CmpltEpi32Mask 

func CmpltEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmplt_epi32_mask'. Requires AVX512F.

func CmpltEpi64Mask 

func CmpltEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmplt_epi64_mask'. Requires AVX512F.

func CmpltEpu32Mask 

func CmpltEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpltEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmplt_epu32_mask'. Requires AVX512F.

func CmpltEpu64Mask 

func CmpltEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmplt_epu64_mask'. Requires AVX512F.

func CmpneqEpi32Mask 

func CmpneqEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpneqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_cmpneq_epi32_mask'. Requires AVX512F.

func CmpneqEpi64Mask 

func CmpneqEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_cmpneq_epi64_mask'. Requires AVX512F.

func CmpneqEpu32Mask 

func CmpneqEpu32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpneqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*32
	k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_cmpneq_epu32_mask'. Requires AVX512F.

func CmpneqEpu64Mask 

func CmpneqEpu64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

CmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 1
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_cmpneq_epu64_mask'. Requires AVX512F.

func ComiRoundSd 

func ComiRoundSd(a x86.M128d, b x86.M128d, imm8 byte, sae int) int

ComiRoundSd: Compare the lower double-precision (64-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and return the boolean result (0 or 1). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	RETURN ( a[63:0] OP b[63:0] ) ? 1 : 0

Instruction: 'VCOMISD'. Intrinsic: '_mm_comi_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func ComiRoundSs 

func ComiRoundSs(a x86.M128, b x86.M128, imm8 byte, sae int) int

ComiRoundSs: Compare the lower single-precision (32-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and return the boolean result (0 or 1). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	RETURN ( a[31:0] OP b[31:0] ) ? 1 : 0

Instruction: 'VCOMISS'. Intrinsic: '_mm_comi_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func CvtRoundi32Ss 

func CvtRoundi32Ss(a x86.M128, b int, rounding int) (dst x86.M128)

CvtRoundi32Ss: Convert the 32-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_Int32_To_FP32(b[31:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvt_roundi32_ss'. Requires AVX512F.

func CvtRoundi64Sd 

func CvtRoundi64Sd(a x86.M128d, b int64, rounding int) (dst x86.M128d)

CvtRoundi64Sd: Convert the 64-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_Int64_To_FP64(b[63:0])
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SD'. Intrinsic: '_mm_cvt_roundi64_sd'. Requires AVX512F.

func CvtRoundi64Ss 

func CvtRoundi64Ss(a x86.M128, b int64, rounding int) (dst x86.M128)

CvtRoundi64Ss: Convert the 64-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_Int64_To_FP32(b[63:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvt_roundi64_ss'. Requires AVX512F.

func CvtRoundsdI32 

func CvtRoundsdI32(a x86.M128d, rounding int) int

CvtRoundsdI32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_Int32(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvt_roundsd_i32'. Requires AVX512F.

func CvtRoundsdI64 

func CvtRoundsdI64(a x86.M128d, rounding int) int64

CvtRoundsdI64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_Int64(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvt_roundsd_i64'. Requires AVX512F.

func CvtRoundsdSi32 

func CvtRoundsdSi32(a x86.M128d, rounding int) int

CvtRoundsdSi32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_Int32(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvt_roundsd_si32'. Requires AVX512F.

func CvtRoundsdSi64 

func CvtRoundsdSi64(a x86.M128d, rounding int) int64

CvtRoundsdSi64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_Int64(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvt_roundsd_si64'. Requires AVX512F.

func CvtRoundsdSs 

func CvtRoundsdSs(a x86.M128, b x86.M128d, rounding int) (dst x86.M128)

CvtRoundsdSs: Convert the lower double-precision (64-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_FP32(b[63:0])
		dst[127:32] := a[127:31]
		dst[MAX:64] := 0

Instruction: 'VCVTSD2SS'. Intrinsic: '_mm_cvt_roundsd_ss'. Requires AVX512F.

func CvtRoundsdU32 

func CvtRoundsdU32(a x86.M128d, rounding int) uint32

CvtRoundsdU32: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_UnsignedInt32(a[63:0])

Instruction: 'VCVTSD2USI'. Intrinsic: '_mm_cvt_roundsd_u32'. Requires AVX512F.

func CvtRoundsdU64 

func CvtRoundsdU64(a x86.M128d, rounding int) uint64

CvtRoundsdU64: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_UnsignedInt64(a[63:0])

Instruction: 'VCVTSD2USI'. Intrinsic: '_mm_cvt_roundsd_u64'. Requires AVX512F.

func CvtRoundsi32Ss 

func CvtRoundsi32Ss(a x86.M128, b int, rounding int) (dst x86.M128)

CvtRoundsi32Ss: Convert the 32-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_Int32_To_FP32(b[31:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvt_roundsi32_ss'. Requires AVX512F.

func CvtRoundsi64Sd 

func CvtRoundsi64Sd(a x86.M128d, b int64, rounding int) (dst x86.M128d)

CvtRoundsi64Sd: Convert the 64-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_Int64_To_FP64(b[63:0])
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SD'. Intrinsic: '_mm_cvt_roundsi64_sd'. Requires AVX512F.

func CvtRoundsi64Ss 

func CvtRoundsi64Ss(a x86.M128, b int64, rounding int) (dst x86.M128)

CvtRoundsi64Ss: Convert the 64-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_Int64_To_FP32(b[63:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvt_roundsi64_ss'. Requires AVX512F.

func CvtRoundssI32 

func CvtRoundssI32(a x86.M128, rounding int) int

CvtRoundssI32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_Int32(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvt_roundss_i32'. Requires AVX512F.

func CvtRoundssI64 

func CvtRoundssI64(a x86.M128, rounding int) int64

CvtRoundssI64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_Int64(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvt_roundss_i64'. Requires AVX512F.

func CvtRoundssSd 

func CvtRoundssSd(a x86.M128d, b x86.M128, rounding int) (dst x86.M128d)

CvtRoundssSd: Convert the lower single-precision (32-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_FP64(b[31:0])
		dst[127:64] := a[127:64]
		dst[MAX:64] := 0

Instruction: 'VCVTSS2SD'. Intrinsic: '_mm_cvt_roundss_sd'. Requires AVX512F.

func CvtRoundssSi32 

func CvtRoundssSi32(a x86.M128, rounding int) int

CvtRoundssSi32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_Int32(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvt_roundss_si32'. Requires AVX512F.

func CvtRoundssSi64 

func CvtRoundssSi64(a x86.M128, rounding int) int64

CvtRoundssSi64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_Int64(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvt_roundss_si64'. Requires AVX512F.

func CvtRoundssU32 

func CvtRoundssU32(a x86.M128, rounding int) uint32

CvtRoundssU32: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 32-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_UnsignedInt32(a[31:0])

Instruction: 'VCVTSS2USI'. Intrinsic: '_mm_cvt_roundss_u32'. Requires AVX512F.

func CvtRoundssU64 

func CvtRoundssU64(a x86.M128, rounding int) uint64

CvtRoundssU64: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 64-bit integer, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_UnsignedInt64(a[31:0])

Instruction: 'VCVTSS2USI'. Intrinsic: '_mm_cvt_roundss_u64'. Requires AVX512F.

func CvtRoundu32Ss 

func CvtRoundu32Ss(a x86.M128, b uint32, rounding int) (dst x86.M128)

CvtRoundu32Ss: Convert the unsigned 32-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_UnsignedInt32_To_FP32(b[31:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTUSI2SS'. Intrinsic: '_mm_cvt_roundu32_ss'. Requires AVX512F.

func CvtRoundu64Sd 

func CvtRoundu64Sd(a x86.M128d, b uint64, rounding int) (dst x86.M128d)

CvtRoundu64Sd: Convert the unsigned 64-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_UnsignedInt64_To_FP64(b[63:0])
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VCVTUSI2SD'. Intrinsic: '_mm_cvt_roundu64_sd'. Requires AVX512F.

func CvtRoundu64Ss 

func CvtRoundu64Ss(a x86.M128, b uint64, rounding int) (dst x86.M128)

CvtRoundu64Ss: Convert the unsigned 64-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_UnsignedInt64_To_FP32(b[63:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VCVTUSI2SS'. Intrinsic: '_mm_cvt_roundu64_ss'. Requires AVX512F.

func Cvtepi32Epi16 

func Cvtepi32Epi16(a x86.M128i) (dst x86.M128i)

Cvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 16*j
	dst[k+15:k] := Truncate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm_cvtepi32_epi16'. Requires AVX512F.

func Cvtepi32Epi8 

func Cvtepi32Epi8(a x86.M128i) (dst x86.M128i)

Cvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 8*j
	dst[k+7:k] := Truncate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm_cvtepi32_epi8'. Requires AVX512F.

func Cvtepi64Epi16 

func Cvtepi64Epi16(a x86.M128i) (dst x86.M128i)

Cvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 16*j
	dst[k+15:k] := Truncate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm_cvtepi64_epi16'. Requires AVX512F.

func Cvtepi64Epi32 

func Cvtepi64Epi32(a x86.M128i) (dst x86.M128i)

Cvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 32*j
	dst[k+31:k] := Truncate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm_cvtepi64_epi32'. Requires AVX512F.

func Cvtepi64Epi8 

func Cvtepi64Epi8(a x86.M128i) (dst x86.M128i)

Cvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 8*j
	dst[k+7:k] := Truncate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm_cvtepi64_epi8'. Requires AVX512F.

func Cvtepu32Pd 

func Cvtepu32Pd(a x86.M128i) (dst x86.M128d)

Cvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	l := j*32
	dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm_cvtepu32_pd'. Requires AVX512F.

func Cvti32Sd 

func Cvti32Sd(a x86.M128d, b int) (dst x86.M128d)

Cvti32Sd: Convert the 32-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

dst[63:0] := Convert_Int32_To_FP64(b[31:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VCVTSI2SD'. Intrinsic: '_mm_cvti32_sd'. Requires AVX512F.

func Cvti32Ss 

func Cvti32Ss(a x86.M128, b int) (dst x86.M128)

Cvti32Ss: Convert the 32-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

dst[31:0] := Convert_Int32_To_FP32(b[31:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvti32_ss'. Requires AVX512F.

func Cvti64Sd 

func Cvti64Sd(a x86.M128d, b int64) (dst x86.M128d)

Cvti64Sd: Convert the 64-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

dst[63:0] := Convert_Int64_To_FP64(b[63:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VCVTSI2SD'. Intrinsic: '_mm_cvti64_sd'. Requires AVX512F.

func Cvti64Ss 

func Cvti64Ss(a x86.M128, b int64) (dst x86.M128)

Cvti64Ss: Convert the 64-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

dst[31:0] := Convert_Int64_To_FP32(b[63:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VCVTSI2SS'. Intrinsic: '_mm_cvti64_ss'. Requires AVX512F.

func CvtpdEpu32 

func CvtpdEpu32(a x86.M128d) (dst x86.M128i)

CvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[k+63:k])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm_cvtpd_epu32'. Requires AVX512F.

func CvtpsEpu32 

func CvtpsEpu32(a x86.M128) (dst x86.M128i)

CvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm_cvtps_epu32'. Requires AVX512F.

func CvtsdI32 

func CvtsdI32(a x86.M128d) int

CvtsdI32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

dst[31:0] := Convert_FP64_To_Int32(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvtsd_i32'. Requires AVX512F.

func CvtsdI64 

func CvtsdI64(a x86.M128d) int64

CvtsdI64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

dst[63:0] := Convert_FP64_To_Int64(a[63:0])

Instruction: 'VCVTSD2SI'. Intrinsic: '_mm_cvtsd_i64'. Requires AVX512F.

func CvtsdU32 

func CvtsdU32(a x86.M128d) uint32

CvtsdU32: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 32-bit integer, and store the result in 'dst'. 

dst[31:0] := Convert_FP64_To_UnsignedInt32(a[63:0])

Instruction: 'VCVTSD2USI'. Intrinsic: '_mm_cvtsd_u32'. Requires AVX512F.

func CvtsdU64 

func CvtsdU64(a x86.M128d) uint64

CvtsdU64: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 64-bit integer, and store the result in 'dst'. 

dst[63:0] := Convert_FP64_To_UnsignedInt64(a[63:0])

Instruction: 'VCVTSD2USI'. Intrinsic: '_mm_cvtsd_u64'. Requires AVX512F.

func Cvtsepi32Epi16 

func Cvtsepi32Epi16(a x86.M128i) (dst x86.M128i)

Cvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm_cvtsepi32_epi16'. Requires AVX512F.

func Cvtsepi32Epi8 

func Cvtsepi32Epi8(a x86.M128i) (dst x86.M128i)

Cvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm_cvtsepi32_epi8'. Requires AVX512F.

func Cvtsepi64Epi16 

func Cvtsepi64Epi16(a x86.M128i) (dst x86.M128i)

Cvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm_cvtsepi64_epi16'. Requires AVX512F.

func Cvtsepi64Epi32 

func Cvtsepi64Epi32(a x86.M128i) (dst x86.M128i)

Cvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm_cvtsepi64_epi32'. Requires AVX512F.

func Cvtsepi64Epi8 

func Cvtsepi64Epi8(a x86.M128i) (dst x86.M128i)

Cvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm_cvtsepi64_epi8'. Requires AVX512F.

func CvtssI32 

func CvtssI32(a x86.M128) int

CvtssI32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer, and store the result in 'dst'. 

dst[31:0] := Convert_FP32_To_Int32(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvtss_i32'. Requires AVX512F.

func CvtssI64 

func CvtssI64(a x86.M128) int64

CvtssI64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer, and store the result in 'dst'. 

dst[63:0] := Convert_FP32_To_Int64(a[31:0])

Instruction: 'VCVTSS2SI'. Intrinsic: '_mm_cvtss_i64'. Requires AVX512F.

func CvtssU32 

func CvtssU32(a x86.M128) uint32

CvtssU32: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 32-bit integer, and store the result in 'dst'. 

dst[31:0] := Convert_FP32_To_UnsignedInt32(a[31:0])

Instruction: 'VCVTSS2USI'. Intrinsic: '_mm_cvtss_u32'. Requires AVX512F.

func CvtssU64 

func CvtssU64(a x86.M128) uint64

CvtssU64: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 64-bit integer, and store the result in 'dst'. 

dst[63:0] := Convert_FP32_To_UnsignedInt64(a[31:0])

Instruction: 'VCVTSS2USI'. Intrinsic: '_mm_cvtss_u64'. Requires AVX512F.

func CvttRoundsdI32 

func CvttRoundsdI32(a x86.M128d, rounding int) int

CvttRoundsdI32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_Int32_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvtt_roundsd_i32'. Requires AVX512F.

func CvttRoundsdI64 

func CvttRoundsdI64(a x86.M128d, rounding int) int64

CvttRoundsdI64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_Int64_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvtt_roundsd_i64'. Requires AVX512F.

func CvttRoundsdSi32 

func CvttRoundsdSi32(a x86.M128d, rounding int) int

CvttRoundsdSi32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_Int32_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvtt_roundsd_si32'. Requires AVX512F.

func CvttRoundsdSi64 

func CvttRoundsdSi64(a x86.M128d, rounding int) int64

CvttRoundsdSi64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_Int64_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvtt_roundsd_si64'. Requires AVX512F.

func CvttRoundsdU32 

func CvttRoundsdU32(a x86.M128d, rounding int) uint32

CvttRoundsdU32: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP64_To_UnsignedInt32_Truncate(a[63:0])

Instruction: 'VCVTTSD2USI'. Intrinsic: '_mm_cvtt_roundsd_u32'. Requires AVX512F.

func CvttRoundsdU64 

func CvttRoundsdU64(a x86.M128d, rounding int) uint64

CvttRoundsdU64: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP64_To_UnsignedInt64_Truncate(a[63:0])

Instruction: 'VCVTTSD2USI'. Intrinsic: '_mm_cvtt_roundsd_u64'. Requires AVX512F.

func CvttRoundssI32 

func CvttRoundssI32(a x86.M128, rounding int) int

CvttRoundssI32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_Int32_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvtt_roundss_i32'. Requires AVX512F.

func CvttRoundssI64 

func CvttRoundssI64(a x86.M128, rounding int) int64

CvttRoundssI64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_Int64_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvtt_roundss_i64'. Requires AVX512F.

func CvttRoundssSi32 

func CvttRoundssSi32(a x86.M128, rounding int) int

CvttRoundssSi32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_Int32_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvtt_roundss_si32'. Requires AVX512F.

func CvttRoundssSi64 

func CvttRoundssSi64(a x86.M128, rounding int) int64

CvttRoundssSi64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_Int64_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvtt_roundss_si64'. Requires AVX512F.

func CvttRoundssU32 

func CvttRoundssU32(a x86.M128, rounding int) uint32

CvttRoundssU32: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 32-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := Convert_FP32_To_UnsignedInt32_Truncate(a[31:0])

Instruction: 'VCVTTSS2USI'. Intrinsic: '_mm_cvtt_roundss_u32'. Requires AVX512F.

func CvttRoundssU64 

func CvttRoundssU64(a x86.M128, rounding int) uint64

CvttRoundssU64: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 64-bit integer with truncation, and store the result in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := Convert_FP32_To_UnsignedInt64_Truncate(a[31:0])

Instruction: 'VCVTTSS2USI'. Intrinsic: '_mm_cvtt_roundss_u64'. Requires AVX512F.

func CvttpdEpu32 

func CvttpdEpu32(a x86.M128d) (dst x86.M128i)

CvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[k+63:k])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm_cvttpd_epu32'. Requires AVX512F.

func CvttpsEpu32 

func CvttpsEpu32(a x86.M128) (dst x86.M128i)

CvttpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32_Truncate(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm_cvttps_epu32'. Requires AVX512F.

func CvttsdI32 

func CvttsdI32(a x86.M128d) int

CvttsdI32: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

dst[31:0] := Convert_FP64_To_Int32_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvttsd_i32'. Requires AVX512F.

func CvttsdI64 

func CvttsdI64(a x86.M128d) int64

CvttsdI64: Convert the lower double-precision (64-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

dst[63:0] := Convert_FP64_To_Int64_Truncate(a[63:0])

Instruction: 'VCVTTSD2SI'. Intrinsic: '_mm_cvttsd_i64'. Requires AVX512F.

func CvttsdU32 

func CvttsdU32(a x86.M128d) uint32

CvttsdU32: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 32-bit integer with truncation, and store the result in 'dst'. 

dst[31:0] := Convert_FP64_To_UnsignedInt32_Truncate(a[63:0])

Instruction: 'VCVTTSD2USI'. Intrinsic: '_mm_cvttsd_u32'. Requires AVX512F.

func CvttsdU64 

func CvttsdU64(a x86.M128d) uint64

CvttsdU64: Convert the lower double-precision (64-bit) floating-point element in 'a' to an unsigned 64-bit integer with truncation, and store the result in 'dst'. 

dst[63:0] := Convert_FP64_To_UnsignedInt64_Truncate(a[63:0])

Instruction: 'VCVTTSD2USI'. Intrinsic: '_mm_cvttsd_u64'. Requires AVX512F.

func CvttssI32 

func CvttssI32(a x86.M128) int

CvttssI32: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 32-bit integer with truncation, and store the result in 'dst'. 

dst[31:0] := Convert_FP32_To_Int32_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvttss_i32'. Requires AVX512F.

func CvttssI64 

func CvttssI64(a x86.M128) int64

CvttssI64: Convert the lower single-precision (32-bit) floating-point element in 'a' to a 64-bit integer with truncation, and store the result in 'dst'. 

dst[63:0] := Convert_FP32_To_Int64_Truncate(a[31:0])

Instruction: 'VCVTTSS2SI'. Intrinsic: '_mm_cvttss_i64'. Requires AVX512F.

func CvttssU32 

func CvttssU32(a x86.M128) uint32

CvttssU32: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 32-bit integer with truncation, and store the result in 'dst'. 

dst[31:0] := Convert_FP32_To_UnsignedInt32_Truncate(a[31:0])

Instruction: 'VCVTTSS2USI'. Intrinsic: '_mm_cvttss_u32'. Requires AVX512F.

func CvttssU64 

func CvttssU64(a x86.M128) uint64

CvttssU64: Convert the lower single-precision (32-bit) floating-point element in 'a' to an unsigned 64-bit integer with truncation, and store the result in 'dst'. 

dst[63:0] := Convert_FP32_To_UnsignedInt64_Truncate(a[31:0])

Instruction: 'VCVTTSS2USI'. Intrinsic: '_mm_cvttss_u64'. Requires AVX512F.

func Cvtu32Sd 

func Cvtu32Sd(a x86.M128d, b uint32) (dst x86.M128d)

Cvtu32Sd: Convert the unsigned 32-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

dst[63:0] := Convert_UnsignedInt32_To_FP64(b[31:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VCVTUSI2SD'. Intrinsic: '_mm_cvtu32_sd'. Requires AVX512F.

func Cvtu32Ss 

func Cvtu32Ss(a x86.M128, b uint32) (dst x86.M128)

Cvtu32Ss: Convert the unsigned 32-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

dst[31:0] := Convert_UnsignedInt32_To_FP32(b[31:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VCVTUSI2SS'. Intrinsic: '_mm_cvtu32_ss'. Requires AVX512F.

func Cvtu64Sd 

func Cvtu64Sd(a x86.M128d, b uint64) (dst x86.M128d)

Cvtu64Sd: Convert the unsigned 64-bit integer 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

dst[63:0] := Convert_UnsignedInt64_To_FP64(b[63:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VCVTUSI2SD'. Intrinsic: '_mm_cvtu64_sd'. Requires AVX512F.

func Cvtu64Ss 

func Cvtu64Ss(a x86.M128, b uint64) (dst x86.M128)

Cvtu64Ss: Convert the unsigned 64-bit integer 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

dst[31:0] := Convert_UnsignedInt64_To_FP32(b[63:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VCVTUSI2SS'. Intrinsic: '_mm_cvtu64_ss'. Requires AVX512F.

func Cvtusepi32Epi16 

func Cvtusepi32Epi16(a x86.M128i) (dst x86.M128i)

Cvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm_cvtusepi32_epi16'. Requires AVX512F.

func Cvtusepi32Epi8 

func Cvtusepi32Epi8(a x86.M128i) (dst x86.M128i)

Cvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm_cvtusepi32_epi8'. Requires AVX512F.

func Cvtusepi64Epi16 

func Cvtusepi64Epi16(a x86.M128i) (dst x86.M128i)

Cvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm_cvtusepi64_epi16'. Requires AVX512F.

func Cvtusepi64Epi32 

func Cvtusepi64Epi32(a x86.M128i) (dst x86.M128i)

Cvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm_cvtusepi64_epi32'. Requires AVX512F.

func Cvtusepi64Epi8 

func Cvtusepi64Epi8(a x86.M128i) (dst x86.M128i)

Cvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 1
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm_cvtusepi64_epi8'. Requires AVX512F.

func DivRoundSd 

func DivRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

DivRoundSd: Divide the lower double-precision (64-bit) floating-point element in 'a' by the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := a[63:0] / b[63:0]
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VDIVSD'. Intrinsic: '_mm_div_round_sd'. Requires AVX512F.

func DivRoundSs 

func DivRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

DivRoundSs: Divide the lower single-precision (32-bit) floating-point element in 'a' by the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := a[31:0] / b[31:0]
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VDIVSS'. Intrinsic: '_mm_div_round_ss'. Requires AVX512F.

func FixupimmPd 

func FixupimmPd(a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

FixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func FixupimmPs 

func FixupimmPs(a x86.M128, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

FixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func FixupimmRoundSd 

func FixupimmRoundSd(a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte, rounding int) (dst x86.M128d)

FixupimmRoundSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_fixupimm_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func FixupimmRoundSs 

func FixupimmRoundSs(a x86.M128, b x86.M128, c x86.M128i, imm8 byte, rounding int) (dst x86.M128)

FixupimmRoundSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_fixupimm_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func FixupimmSd 

func FixupimmSd(a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

FixupimmSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_fixupimm_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func FixupimmSs 

func FixupimmSs(a x86.M128, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

FixupimmSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_fixupimm_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func GetexpPd 

func GetexpPd(a x86.M128d) (dst x86.M128d)

GetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := ConvertExpFP64(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm_getexp_pd'. Requires AVX512F.

func GetexpPs 

func GetexpPs(a x86.M128) (dst x86.M128)

GetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := ConvertExpFP32(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm_getexp_ps'. Requires AVX512F.

func GetexpRoundSd 

func GetexpRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

GetexpRoundSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := ConvertExpFP64(b[63:0])
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_getexp_round_sd'. Requires AVX512F.

func GetexpRoundSs 

func GetexpRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

GetexpRoundSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := ConvertExpFP32(b[31:0])
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_getexp_round_ss'. Requires AVX512F.

func GetexpSd 

func GetexpSd(a x86.M128d, b x86.M128d) (dst x86.M128d)

GetexpSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

dst[63:0] := ConvertExpFP64(b[63:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_getexp_sd'. Requires AVX512F.

func GetexpSs 

func GetexpSs(a x86.M128, b x86.M128) (dst x86.M128)

GetexpSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

dst[31:0] := ConvertExpFP32(b[31:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_getexp_ss'. Requires AVX512F.

func GetmantPd 

func GetmantPd(a x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

GetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 1
			i := j*64
			dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm_getmant_pd'. Requires AVX512F.

func GetmantPs 

func GetmantPs(a x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

GetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*32
			dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm_getmant_ps'. Requires AVX512F.

func GetmantRoundSd 

func GetmantRoundSd(a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128d)

GetmantRoundSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_getmant_round_sd'. Requires AVX512F.

func GetmantRoundSs 

func GetmantRoundSs(a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128)

GetmantRoundSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_getmant_round_ss'. Requires AVX512F.

func GetmantSd 

func GetmantSd(a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

GetmantSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_getmant_sd'. Requires AVX512F.

func GetmantSs 

func GetmantSs(a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

GetmantSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_getmant_ss'. Requires AVX512F.

func M256AbsEpi64 

func M256AbsEpi64(a x86.M256i) (dst x86.M256i)

M256AbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := ABS(a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm256_abs_epi64'. Requires AVX512F.

func M256BroadcastF32x4 

func M256BroadcastF32x4(a x86.M128) (dst x86.M256)

M256BroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	dst[i+31:i] := a[n+31:n]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm256_broadcast_f32x4'. Requires AVX512F.

func M256BroadcastI32x4 

func M256BroadcastI32x4(a x86.M128i) (dst x86.M256i)

M256BroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	dst[i+31:i] := a[n+31:n]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm256_broadcast_i32x4'. Requires AVX512F.

func M256CmpEpi32Mask 

func M256CmpEpi32Mask(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256CmpEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmp_epi32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpEpi64Mask 

func M256CmpEpi64Mask(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256CmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpEpu32Mask 

func M256CmpEpu32Mask(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256CmpEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmp_epu32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpEpu64Mask 

func M256CmpEpu64Mask(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256CmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpPdMask 

func M256CmpPdMask(a x86.M256d, b x86.M256d, imm8 byte) (dst x86.Mmask8)

M256CmpPdMask: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 3
	i := j*64
	k[j] := (a[i+63:i] OP b[i+63:i]) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VCMPPD'. Intrinsic: '_mm256_cmp_pd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpPsMask 

func M256CmpPsMask(a x86.M256, b x86.M256, imm8 byte) (dst x86.Mmask8)

M256CmpPsMask: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 7
	i := j*32
	k[j] := (a[i+31:i] OP b[i+31:i]) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VCMPPS'. Intrinsic: '_mm256_cmp_ps_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256CmpeqEpi32Mask 

func M256CmpeqEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpeqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmpeq_epi32_mask'. Requires AVX512F.

func M256CmpeqEpi64Mask 

func M256CmpeqEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmpeq_epi64_mask'. Requires AVX512F.

func M256CmpeqEpu32Mask 

func M256CmpeqEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpeqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmpeq_epu32_mask'. Requires AVX512F.

func M256CmpeqEpu64Mask 

func M256CmpeqEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmpeq_epu64_mask'. Requires AVX512F.

func M256CmpgeEpi32Mask 

func M256CmpgeEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgeEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmpge_epi32_mask'. Requires AVX512F.

func M256CmpgeEpi64Mask 

func M256CmpgeEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmpge_epi64_mask'. Requires AVX512F.

func M256CmpgeEpu32Mask 

func M256CmpgeEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgeEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmpge_epu32_mask'. Requires AVX512F.

func M256CmpgeEpu64Mask 

func M256CmpgeEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmpge_epu64_mask'. Requires AVX512F.

func M256CmpgtEpi32Mask 

func M256CmpgtEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgtEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmpgt_epi32_mask'. Requires AVX512F.

func M256CmpgtEpi64Mask 

func M256CmpgtEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmpgt_epi64_mask'. Requires AVX512F.

func M256CmpgtEpu32Mask 

func M256CmpgtEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgtEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmpgt_epu32_mask'. Requires AVX512F.

func M256CmpgtEpu64Mask 

func M256CmpgtEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmpgt_epu64_mask'. Requires AVX512F.

func M256CmpleEpi32Mask 

func M256CmpleEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpleEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmple_epi32_mask'. Requires AVX512F.

func M256CmpleEpi64Mask 

func M256CmpleEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmple_epi64_mask'. Requires AVX512F.

func M256CmpleEpu32Mask 

func M256CmpleEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpleEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmple_epu32_mask'. Requires AVX512F.

func M256CmpleEpu64Mask 

func M256CmpleEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmple_epu64_mask'. Requires AVX512F.

func M256CmpltEpi32Mask 

func M256CmpltEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmplt_epi32_mask'. Requires AVX512F.

func M256CmpltEpi64Mask 

func M256CmpltEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmplt_epi64_mask'. Requires AVX512F.

func M256CmpltEpu32Mask 

func M256CmpltEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpltEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmplt_epu32_mask'. Requires AVX512F.

func M256CmpltEpu64Mask 

func M256CmpltEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmplt_epu64_mask'. Requires AVX512F.

func M256CmpneqEpi32Mask 

func M256CmpneqEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpneqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_cmpneq_epi32_mask'. Requires AVX512F.

func M256CmpneqEpi64Mask 

func M256CmpneqEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_cmpneq_epi64_mask'. Requires AVX512F.

func M256CmpneqEpu32Mask 

func M256CmpneqEpu32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpneqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*32
	k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_cmpneq_epu32_mask'. Requires AVX512F.

func M256CmpneqEpu64Mask 

func M256CmpneqEpu64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256CmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 3
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_cmpneq_epu64_mask'. Requires AVX512F.

func M256Cvtepi32Epi16 

func M256Cvtepi32Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 16*j
	dst[k+15:k] := Truncate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm256_cvtepi32_epi16'. Requires AVX512F.

func M256Cvtepi32Epi8 

func M256Cvtepi32Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 8*j
	dst[k+7:k] := Truncate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm256_cvtepi32_epi8'. Requires AVX512F.

func M256Cvtepi64Epi16 

func M256Cvtepi64Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 16*j
	dst[k+15:k] := Truncate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm256_cvtepi64_epi16'. Requires AVX512F.

func M256Cvtepi64Epi32 

func M256Cvtepi64Epi32(a x86.M256i) (dst x86.M128i)

M256Cvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 32*j
	dst[k+31:k] := Truncate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm256_cvtepi64_epi32'. Requires AVX512F.

func M256Cvtepi64Epi8 

func M256Cvtepi64Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 8*j
	dst[k+7:k] := Truncate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm256_cvtepi64_epi8'. Requires AVX512F.

func M256Cvtepu32Pd 

func M256Cvtepu32Pd(a x86.M128i) (dst x86.M256d)

M256Cvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	l := j*32
	dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm256_cvtepu32_pd'. Requires AVX512F.

func M256CvtpdEpu32 

func M256CvtpdEpu32(a x86.M256d) (dst x86.M128i)

M256CvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[k+63:k])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm256_cvtpd_epu32'. Requires AVX512F.

func M256CvtpsEpu32 

func M256CvtpsEpu32(a x86.M256) (dst x86.M256i)

M256CvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm256_cvtps_epu32'. Requires AVX512F.

func M256Cvtsepi32Epi16 

func M256Cvtsepi32Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm256_cvtsepi32_epi16'. Requires AVX512F.

func M256Cvtsepi32Epi8 

func M256Cvtsepi32Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm256_cvtsepi32_epi8'. Requires AVX512F.

func M256Cvtsepi64Epi16 

func M256Cvtsepi64Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm256_cvtsepi64_epi16'. Requires AVX512F.

func M256Cvtsepi64Epi32 

func M256Cvtsepi64Epi32(a x86.M256i) (dst x86.M128i)

M256Cvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm256_cvtsepi64_epi32'. Requires AVX512F.

func M256Cvtsepi64Epi8 

func M256Cvtsepi64Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm256_cvtsepi64_epi8'. Requires AVX512F.

func M256CvttpdEpu32 

func M256CvttpdEpu32(a x86.M256d) (dst x86.M128i)

M256CvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[k+63:k])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm256_cvttpd_epu32'. Requires AVX512F.

func M256CvttpsEpu32 

func M256CvttpsEpu32(a x86.M256) (dst x86.M256i)

M256CvttpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32_Truncate(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm256_cvttps_epu32'. Requires AVX512F.

func M256Cvtusepi32Epi16 

func M256Cvtusepi32Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm256_cvtusepi32_epi16'. Requires AVX512F.

func M256Cvtusepi32Epi8 

func M256Cvtusepi32Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm256_cvtusepi32_epi8'. Requires AVX512F.

func M256Cvtusepi64Epi16 

func M256Cvtusepi64Epi16(a x86.M256i) (dst x86.M128i)

M256Cvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm256_cvtusepi64_epi16'. Requires AVX512F.

func M256Cvtusepi64Epi32 

func M256Cvtusepi64Epi32(a x86.M256i) (dst x86.M128i)

M256Cvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm256_cvtusepi64_epi32'. Requires AVX512F.

func M256Cvtusepi64Epi8 

func M256Cvtusepi64Epi8(a x86.M256i) (dst x86.M128i)

M256Cvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 3
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm256_cvtusepi64_epi8'. Requires AVX512F.

func M256Extractf32x4Ps 

func M256Extractf32x4Ps(a x86.M256, imm8 byte) (dst x86.M128)

M256Extractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
ESAC
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm256_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256Extracti32x4Epi32 

func M256Extracti32x4Epi32(a x86.M256i, imm8 byte) (dst x86.M128i)

M256Extracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
ESAC
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm256_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256FixupimmPd 

func M256FixupimmPd(a x86.M256d, b x86.M256d, c x86.M256i, imm8 byte) (dst x86.M256d)

M256FixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm256_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256FixupimmPs 

func M256FixupimmPs(a x86.M256, b x86.M256, c x86.M256i, imm8 byte) (dst x86.M256)

M256FixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm256_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256GetexpPd 

func M256GetexpPd(a x86.M256d) (dst x86.M256d)

M256GetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := ConvertExpFP64(a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm256_getexp_pd'. Requires AVX512F.

func M256GetexpPs 

func M256GetexpPs(a x86.M256) (dst x86.M256)

M256GetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := ConvertExpFP32(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm256_getexp_ps'. Requires AVX512F.

func M256GetmantPd 

func M256GetmantPd(a x86.M256d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256d)

M256GetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*64
			dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm256_getmant_pd'. Requires AVX512F.

func M256GetmantPs 

func M256GetmantPs(a x86.M256, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256)

M256GetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 7
			i := j*32
			dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm256_getmant_ps'. Requires AVX512F.

func M256Insertf32x4 

func M256Insertf32x4(a x86.M256, b x86.M128, imm8 byte) (dst x86.M256)

M256Insertf32x4: Copy 'a' to 'dst', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: dst[127:0] := b[127:0]
1: dst[255:128] := b[127:0]
ESAC
dst[MAX:256] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm256_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256Inserti32x4 

func M256Inserti32x4(a x86.M256i, b x86.M128i, imm8 byte) (dst x86.M256i)

M256Inserti32x4: Copy 'a' to 'dst', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: dst[127:0] := b[127:0]
1: dst[255:128] := b[127:0]
ESAC
dst[MAX:256] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm256_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256Mask2Permutex2varEpi32 

func M256Mask2Permutex2varEpi32(a x86.M256i, idx x86.M256i, k x86.Mmask8, b x86.M256i) (dst x86.M256i)

M256Mask2Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2D'. Intrinsic: '_mm256_mask2_permutex2var_epi32'. Requires AVX512F.

func M256Mask2Permutex2varEpi64 

func M256Mask2Permutex2varEpi64(a x86.M256i, idx x86.M256i, k x86.Mmask8, b x86.M256i) (dst x86.M256i)

M256Mask2Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2Q'. Intrinsic: '_mm256_mask2_permutex2var_epi64'. Requires AVX512F.

func M256Mask2Permutex2varPd 

func M256Mask2Permutex2varPd(a x86.M256d, idx x86.M256i, k x86.Mmask8, b x86.M256d) (dst x86.M256d)

M256Mask2Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PD'. Intrinsic: '_mm256_mask2_permutex2var_pd'. Requires AVX512F.

func M256Mask2Permutex2varPs 

func M256Mask2Permutex2varPs(a x86.M256, idx x86.M256i, k x86.Mmask8, b x86.M256) (dst x86.M256)

M256Mask2Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PS'. Intrinsic: '_mm256_mask2_permutex2var_ps'. Requires AVX512F.

func M256Mask3FmaddPd 

func M256Mask3FmaddPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm256_mask3_fmadd_pd'. Requires AVX512F.

func M256Mask3FmaddPs 

func M256Mask3FmaddPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm256_mask3_fmadd_ps'. Requires AVX512F.

func M256Mask3FmaddsubPd 

func M256Mask3FmaddsubPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm256_mask3_fmaddsub_pd'. Requires AVX512F.

func M256Mask3FmaddsubPs 

func M256Mask3FmaddsubPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm256_mask3_fmaddsub_ps'. Requires AVX512F.

func M256Mask3FmsubPd 

func M256Mask3FmsubPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm256_mask3_fmsub_pd'. Requires AVX512F.

func M256Mask3FmsubPs 

func M256Mask3FmsubPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm256_mask3_fmsub_ps'. Requires AVX512F.

func M256Mask3FmsubaddPd 

func M256Mask3FmsubaddPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm256_mask3_fmsubadd_pd'. Requires AVX512F.

func M256Mask3FmsubaddPs 

func M256Mask3FmsubaddPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm256_mask3_fmsubadd_ps'. Requires AVX512F.

func M256Mask3FnmaddPd 

func M256Mask3FnmaddPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm256_mask3_fnmadd_pd'. Requires AVX512F.

func M256Mask3FnmaddPs 

func M256Mask3FnmaddPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm256_mask3_fnmadd_ps'. Requires AVX512F.

func M256Mask3FnmsubPd 

func M256Mask3FnmsubPd(a x86.M256d, b x86.M256d, c x86.M256d, k x86.Mmask8) (dst x86.M256d)

M256Mask3FnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm256_mask3_fnmsub_pd'. Requires AVX512F.

func M256Mask3FnmsubPs 

func M256Mask3FnmsubPs(a x86.M256, b x86.M256, c x86.M256, k x86.Mmask8) (dst x86.M256)

M256Mask3FnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm256_mask3_fnmsub_ps'. Requires AVX512F.

func M256MaskAbsEpi32 

func M256MaskAbsEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm256_mask_abs_epi32'. Requires AVX512F.

func M256MaskAbsEpi64 

func M256MaskAbsEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm256_mask_abs_epi64'. Requires AVX512F.

func M256MaskAddEpi32 

func M256MaskAddEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAddEpi32: Add packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPADDD'. Intrinsic: '_mm256_mask_add_epi32'. Requires AVX512F.

func M256MaskAddEpi64 

func M256MaskAddEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm256_mask_add_epi64'. Requires AVX512F.

func M256MaskAndEpi32 

func M256MaskAndEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAndEpi32: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] BITWISE AND b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDD'. Intrinsic: '_mm256_mask_and_epi32'. Requires AVX512F.

func M256MaskAndEpi64 

func M256MaskAndEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAndEpi64: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] AND b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDQ'. Intrinsic: '_mm256_mask_and_epi64'. Requires AVX512F.

func M256MaskAndnotEpi32 

func M256MaskAndnotEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAndnotEpi32: Compute the bitwise AND NOT of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ((NOT a[i+31:i]) AND b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDND'. Intrinsic: '_mm256_mask_andnot_epi32'. Requires AVX512F.

func M256MaskAndnotEpi64 

func M256MaskAndnotEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskAndnotEpi64: Compute the bitwise AND NOT of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ((NOT a[i+63:i]) AND b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDNQ'. Intrinsic: '_mm256_mask_andnot_epi64'. Requires AVX512F.

func M256MaskBlendEpi32 

func M256MaskBlendEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskBlendEpi32: Blend packed 32-bit integers from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := b[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBLENDMD'. Intrinsic: '_mm256_mask_blend_epi32'. Requires AVX512F.

func M256MaskBlendEpi64 

func M256MaskBlendEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskBlendEpi64: Blend packed 64-bit integers from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := b[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBLENDMQ'. Intrinsic: '_mm256_mask_blend_epi64'. Requires AVX512F.

func M256MaskBlendPd 

func M256MaskBlendPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskBlendPd: Blend packed double-precision (64-bit) floating-point elements from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := b[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBLENDMPD'. Intrinsic: '_mm256_mask_blend_pd'. Requires AVX512F.

func M256MaskBlendPs 

func M256MaskBlendPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskBlendPs: Blend packed single-precision (32-bit) floating-point elements from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := b[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBLENDMPS'. Intrinsic: '_mm256_mask_blend_ps'. Requires AVX512F.

func M256MaskBroadcastF32x4 

func M256MaskBroadcastF32x4(src x86.M256, k x86.Mmask8, a x86.M128) (dst x86.M256)

M256MaskBroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm256_mask_broadcast_f32x4'. Requires AVX512F.

func M256MaskBroadcastI32x4 

func M256MaskBroadcastI32x4(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskBroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := src[n+31:n]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm256_mask_broadcast_i32x4'. Requires AVX512F.

func M256MaskBroadcastdEpi32 

func M256MaskBroadcastdEpi32(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm256_mask_broadcastd_epi32'. Requires AVX512F.

func M256MaskBroadcastqEpi64 

func M256MaskBroadcastqEpi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm256_mask_broadcastq_epi64'. Requires AVX512F.

func M256MaskBroadcastsdPd 

func M256MaskBroadcastsdPd(src x86.M256d, k x86.Mmask8, a x86.M128d) (dst x86.M256d)

M256MaskBroadcastsdPd: Broadcast the low double-precision (64-bit) floating-point element from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTSD'. Intrinsic: '_mm256_mask_broadcastsd_pd'. Requires AVX512F.

func M256MaskBroadcastssPs 

func M256MaskBroadcastssPs(src x86.M256, k x86.Mmask8, a x86.M128) (dst x86.M256)

M256MaskBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm256_mask_broadcastss_ps'. Requires AVX512F.

func M256MaskCmpEpi32Mask 

func M256MaskCmpEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256MaskCmpEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmp_epi32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpEpi64Mask 

func M256MaskCmpEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256MaskCmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpEpu32Mask 

func M256MaskCmpEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256MaskCmpEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmp_epu32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpEpu64Mask 

func M256MaskCmpEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.Mmask8)

M256MaskCmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpPdMask 

func M256MaskCmpPdMask(k1 x86.Mmask8, a x86.M256d, b x86.M256d, imm8 byte) (dst x86.Mmask8)

M256MaskCmpPdMask: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k' using zeromask 'k1' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VCMPPD'. Intrinsic: '_mm256_mask_cmp_pd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpPsMask 

func M256MaskCmpPsMask(k1 x86.Mmask8, a x86.M256, b x86.M256, imm8 byte) (dst x86.Mmask8)

M256MaskCmpPsMask: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k' using zeromask 'k1' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VCMPPS'. Intrinsic: '_mm256_mask_cmp_ps_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskCmpeqEpi32Mask 

func M256MaskCmpeqEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpeqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmpeq_epi32_mask'. Requires AVX512F.

func M256MaskCmpeqEpi64Mask 

func M256MaskCmpeqEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmpeq_epi64_mask'. Requires AVX512F.

func M256MaskCmpeqEpu32Mask 

func M256MaskCmpeqEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpeqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmpeq_epu32_mask'. Requires AVX512F.

func M256MaskCmpeqEpu64Mask 

func M256MaskCmpeqEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmpeq_epu64_mask'. Requires AVX512F.

func M256MaskCmpgeEpi32Mask 

func M256MaskCmpgeEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgeEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmpge_epi32_mask'. Requires AVX512F.

func M256MaskCmpgeEpi64Mask 

func M256MaskCmpgeEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmpge_epi64_mask'. Requires AVX512F.

func M256MaskCmpgeEpu32Mask 

func M256MaskCmpgeEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgeEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmpge_epu32_mask'. Requires AVX512F.

func M256MaskCmpgeEpu64Mask 

func M256MaskCmpgeEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmpge_epu64_mask'. Requires AVX512F.

func M256MaskCmpgtEpi32Mask 

func M256MaskCmpgtEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgtEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmpgt_epi32_mask'. Requires AVX512F.

func M256MaskCmpgtEpi64Mask 

func M256MaskCmpgtEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmpgt_epi64_mask'. Requires AVX512F.

func M256MaskCmpgtEpu32Mask 

func M256MaskCmpgtEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgtEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmpgt_epu32_mask'. Requires AVX512F.

func M256MaskCmpgtEpu64Mask 

func M256MaskCmpgtEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmpgt_epu64_mask'. Requires AVX512F.

func M256MaskCmpleEpi32Mask 

func M256MaskCmpleEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpleEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmple_epi32_mask'. Requires AVX512F.

func M256MaskCmpleEpi64Mask 

func M256MaskCmpleEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmple_epi64_mask'. Requires AVX512F.

func M256MaskCmpleEpu32Mask 

func M256MaskCmpleEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpleEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmple_epu32_mask'. Requires AVX512F.

func M256MaskCmpleEpu64Mask 

func M256MaskCmpleEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmple_epu64_mask'. Requires AVX512F.

func M256MaskCmpltEpi32Mask 

func M256MaskCmpltEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmplt_epi32_mask'. Requires AVX512F.

func M256MaskCmpltEpi64Mask 

func M256MaskCmpltEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmplt_epi64_mask'. Requires AVX512F.

func M256MaskCmpltEpu32Mask 

func M256MaskCmpltEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpltEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmplt_epu32_mask'. Requires AVX512F.

func M256MaskCmpltEpu64Mask 

func M256MaskCmpltEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmplt_epu64_mask'. Requires AVX512F.

func M256MaskCmpneqEpi32Mask 

func M256MaskCmpneqEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpneqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm256_mask_cmpneq_epi32_mask'. Requires AVX512F.

func M256MaskCmpneqEpi64Mask 

func M256MaskCmpneqEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm256_mask_cmpneq_epi64_mask'. Requires AVX512F.

func M256MaskCmpneqEpu32Mask 

func M256MaskCmpneqEpu32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpneqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm256_mask_cmpneq_epu32_mask'. Requires AVX512F.

func M256MaskCmpneqEpu64Mask 

func M256MaskCmpneqEpu64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskCmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm256_mask_cmpneq_epu64_mask'. Requires AVX512F.

func M256MaskCompressEpi32 

func M256MaskCompressEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := src[255:m]
dst[MAX:256] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm256_mask_compress_epi32'. Requires AVX512F.

func M256MaskCompressEpi64 

func M256MaskCompressEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := src[255:m]
dst[MAX:256] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm256_mask_compress_epi64'. Requires AVX512F.

func M256MaskCompressPd 

func M256MaskCompressPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := src[255:m]
dst[MAX:256] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm256_mask_compress_pd'. Requires AVX512F.

func M256MaskCompressPs 

func M256MaskCompressPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := src[255:m]
dst[MAX:256] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm256_mask_compress_ps'. Requires AVX512F.

func M256MaskCvtRoundpsPh 

func M256MaskCvtRoundpsPh(src x86.M128i, k x86.Mmask8, a x86.M256, rounding int) (dst x86.M128i)

M256MaskCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm256_mask_cvt_roundps_ph'. Requires AVX512F.

func M256MaskCvtepi16Epi32 

func M256MaskCvtepi16Epi32(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	l := j*16
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm256_mask_cvtepi16_epi32'. Requires AVX512F.

func M256MaskCvtepi16Epi64 

func M256MaskCvtepi16Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepi16Epi64: Sign extend packed 16-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm256_mask_cvtepi16_epi64'. Requires AVX512F.

func M256MaskCvtepi32Epi16 

func M256MaskCvtepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm256_mask_cvtepi32_epi16'. Requires AVX512F.

func M256MaskCvtepi32Epi64 

func M256MaskCvtepi32Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm256_mask_cvtepi32_epi64'. Requires AVX512F.

func M256MaskCvtepi32Epi8 

func M256MaskCvtepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm256_mask_cvtepi32_epi8'. Requires AVX512F.

func M256MaskCvtepi32Pd 

func M256MaskCvtepi32Pd(src x86.M256d, k x86.Mmask8, a x86.M128i) (dst x86.M256d)

M256MaskCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := src[m+63:m]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm256_mask_cvtepi32_pd'. Requires AVX512F.

func M256MaskCvtepi32Ps 

func M256MaskCvtepi32Ps(src x86.M256, k x86.Mmask8, a x86.M256i) (dst x86.M256)

M256MaskCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm256_mask_cvtepi32_ps'. Requires AVX512F.

func M256MaskCvtepi64Epi16 

func M256MaskCvtepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm256_mask_cvtepi64_epi16'. Requires AVX512F.

func M256MaskCvtepi64Epi32 

func M256MaskCvtepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm256_mask_cvtepi64_epi32'. Requires AVX512F.

func M256MaskCvtepi64Epi8 

func M256MaskCvtepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm256_mask_cvtepi64_epi8'. Requires AVX512F.

func M256MaskCvtepi8Epi32 

func M256MaskCvtepi8Epi32(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepi8Epi32: Sign extend packed 8-bit integers in the low 8 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm256_mask_cvtepi8_epi32'. Requires AVX512F.

func M256MaskCvtepi8Epi64 

func M256MaskCvtepi8Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepi8Epi64: Sign extend packed 8-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm256_mask_cvtepi8_epi64'. Requires AVX512F.

func M256MaskCvtepu16Epi32 

func M256MaskCvtepu16Epi32(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm256_mask_cvtepu16_epi32'. Requires AVX512F.

func M256MaskCvtepu16Epi64 

func M256MaskCvtepu16Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm256_mask_cvtepu16_epi64'. Requires AVX512F.

func M256MaskCvtepu32Epi64 

func M256MaskCvtepu32Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm256_mask_cvtepu32_epi64'. Requires AVX512F.

func M256MaskCvtepu32Pd 

func M256MaskCvtepu32Pd(src x86.M256d, k x86.Mmask8, a x86.M128i) (dst x86.M256d)

M256MaskCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm256_mask_cvtepu32_pd'. Requires AVX512F.

func M256MaskCvtepu8Epi32 

func M256MaskCvtepu8Epi32(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in the low 8 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm256_mask_cvtepu8_epi32'. Requires AVX512F.

func M256MaskCvtepu8Epi64 

func M256MaskCvtepu8Epi64(src x86.M256i, k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm256_mask_cvtepu8_epi64'. Requires AVX512F.

func M256MaskCvtpdEpi32 

func M256MaskCvtpdEpi32(src x86.M128i, k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm256_mask_cvtpd_epi32'. Requires AVX512F.

func M256MaskCvtpdEpu32 

func M256MaskCvtpdEpu32(src x86.M128i, k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm256_mask_cvtpd_epu32'. Requires AVX512F.

func M256MaskCvtpdPs 

func M256MaskCvtpdPs(src x86.M128, k x86.Mmask8, a x86.M256d) (dst x86.M128)

M256MaskCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm256_mask_cvtpd_ps'. Requires AVX512F.

func M256MaskCvtphPs 

func M256MaskCvtphPs(src x86.M256, k x86.Mmask8, a x86.M128i) (dst x86.M256)

M256MaskCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm256_mask_cvtph_ps'. Requires AVX512F.

func M256MaskCvtpsEpi32 

func M256MaskCvtpsEpi32(src x86.M256i, k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm256_mask_cvtps_epi32'. Requires AVX512F.

func M256MaskCvtpsEpu32 

func M256MaskCvtpsEpu32(src x86.M256i, k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm256_mask_cvtps_epu32'. Requires AVX512F.

func M256MaskCvtpsPh 

func M256MaskCvtpsPh(src x86.M128i, k x86.Mmask8, a x86.M256, rounding int) (dst x86.M128i)

M256MaskCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm256_mask_cvtps_ph'. Requires AVX512F.

func M256MaskCvtsepi32Epi16 

func M256MaskCvtsepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm256_mask_cvtsepi32_epi16'. Requires AVX512F.

func M256MaskCvtsepi32Epi8 

func M256MaskCvtsepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm256_mask_cvtsepi32_epi8'. Requires AVX512F.

func M256MaskCvtsepi64Epi16 

func M256MaskCvtsepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm256_mask_cvtsepi64_epi16'. Requires AVX512F.

func M256MaskCvtsepi64Epi32 

func M256MaskCvtsepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm256_mask_cvtsepi64_epi32'. Requires AVX512F.

func M256MaskCvtsepi64Epi8 

func M256MaskCvtsepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm256_mask_cvtsepi64_epi8'. Requires AVX512F.

func M256MaskCvttpdEpi32 

func M256MaskCvttpdEpi32(src x86.M128i, k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm256_mask_cvttpd_epi32'. Requires AVX512F.

func M256MaskCvttpdEpu32 

func M256MaskCvttpdEpu32(src x86.M128i, k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm256_mask_cvttpd_epu32'. Requires AVX512F.

func M256MaskCvttpsEpi32 

func M256MaskCvttpsEpi32(src x86.M256i, k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm256_mask_cvttps_epi32'. Requires AVX512F.

func M256MaskCvttpsEpu32 

func M256MaskCvttpsEpu32(src x86.M256i, k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm256_mask_cvttps_epu32'. Requires AVX512F.

func M256MaskCvtusepi32Epi16 

func M256MaskCvtusepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm256_mask_cvtusepi32_epi16'. Requires AVX512F.

func M256MaskCvtusepi32Epi8 

func M256MaskCvtusepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm256_mask_cvtusepi32_epi8'. Requires AVX512F.

func M256MaskCvtusepi64Epi16 

func M256MaskCvtusepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm256_mask_cvtusepi64_epi16'. Requires AVX512F.

func M256MaskCvtusepi64Epi32 

func M256MaskCvtusepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm256_mask_cvtusepi64_epi32'. Requires AVX512F.

func M256MaskCvtusepi64Epi8 

func M256MaskCvtusepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm256_mask_cvtusepi64_epi8'. Requires AVX512F.

func M256MaskDivPd 

func M256MaskDivPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm256_mask_div_pd'. Requires AVX512F.

func M256MaskDivPs 

func M256MaskDivPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm256_mask_div_ps'. Requires AVX512F.

func M256MaskExpandEpi32 

func M256MaskExpandEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm256_mask_expand_epi32'. Requires AVX512F.

func M256MaskExpandEpi64 

func M256MaskExpandEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm256_mask_expand_epi64'. Requires AVX512F.

func M256MaskExpandPd 

func M256MaskExpandPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm256_mask_expand_pd'. Requires AVX512F.

func M256MaskExpandPs 

func M256MaskExpandPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm256_mask_expand_ps'. Requires AVX512F.

func M256MaskExtractf32x4Ps 

func M256MaskExtractf32x4Ps(src x86.M128, k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M128)

M256MaskExtractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[127:0] := a[127:0]
1: tmp[127:0] := a[255:128]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm256_mask_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskExtracti32x4Epi32 

func M256MaskExtracti32x4Epi32(src x86.M128i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M128i)

M256MaskExtracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm256_mask_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskFixupimmPd 

func M256MaskFixupimmPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256i, imm8 byte) (dst x86.M256d)

M256MaskFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm256_mask_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskFixupimmPs 

func M256MaskFixupimmPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256i, imm8 byte) (dst x86.M256)

M256MaskFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm256_mask_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskFmaddPd 

func M256MaskFmaddPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm256_mask_fmadd_pd'. Requires AVX512F.

func M256MaskFmaddPs 

func M256MaskFmaddPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm256_mask_fmadd_ps'. Requires AVX512F.

func M256MaskFmaddsubPd 

func M256MaskFmaddsubPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm256_mask_fmaddsub_pd'. Requires AVX512F.

func M256MaskFmaddsubPs 

func M256MaskFmaddsubPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm256_mask_fmaddsub_ps'. Requires AVX512F.

func M256MaskFmsubPd 

func M256MaskFmsubPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm256_mask_fmsub_pd'. Requires AVX512F.

func M256MaskFmsubPs 

func M256MaskFmsubPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm256_mask_fmsub_ps'. Requires AVX512F.

func M256MaskFmsubaddPd 

func M256MaskFmsubaddPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm256_mask_fmsubadd_pd'. Requires AVX512F.

func M256MaskFmsubaddPs 

func M256MaskFmsubaddPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm256_mask_fmsubadd_ps'. Requires AVX512F.

func M256MaskFnmaddPd 

func M256MaskFnmaddPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm256_mask_fnmadd_pd'. Requires AVX512F.

func M256MaskFnmaddPs 

func M256MaskFnmaddPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm256_mask_fnmadd_ps'. Requires AVX512F.

func M256MaskFnmsubPd 

func M256MaskFnmsubPd(a x86.M256d, k x86.Mmask8, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskFnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm256_mask_fnmsub_pd'. Requires AVX512F.

func M256MaskFnmsubPs 

func M256MaskFnmsubPs(a x86.M256, k x86.Mmask8, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskFnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm256_mask_fnmsub_ps'. Requires AVX512F.

func M256MaskGetexpPd 

func M256MaskGetexpPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskGetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm256_mask_getexp_pd'. Requires AVX512F.

func M256MaskGetexpPs 

func M256MaskGetexpPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskGetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm256_mask_getexp_ps'. Requires AVX512F.

func M256MaskGetmantPd 

func M256MaskGetmantPd(src x86.M256d, k x86.Mmask8, a x86.M256d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256d)

M256MaskGetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm256_mask_getmant_pd'. Requires AVX512F.

func M256MaskGetmantPs 

func M256MaskGetmantPs(src x86.M256, k x86.Mmask8, a x86.M256, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256)

M256MaskGetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 7
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm256_mask_getmant_ps'. Requires AVX512F.

func M256MaskInsertf32x4 

func M256MaskInsertf32x4(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M128, imm8 byte) (dst x86.M256)

M256MaskInsertf32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
ESAC
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm256_mask_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskInserti32x4 

func M256MaskInserti32x4(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M128i, imm8 byte) (dst x86.M256i)

M256MaskInserti32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
ESAC
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm256_mask_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskMaxEpi32 

func M256MaskMaxEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMaxEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXSD'. Intrinsic: '_mm256_mask_max_epi32'. Requires AVX512F.

func M256MaskMaxEpi64 

func M256MaskMaxEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm256_mask_max_epi64'. Requires AVX512F.

func M256MaskMaxEpu32 

func M256MaskMaxEpu32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMaxEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXUD'. Intrinsic: '_mm256_mask_max_epu32'. Requires AVX512F.

func M256MaskMaxEpu64 

func M256MaskMaxEpu64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm256_mask_max_epu64'. Requires AVX512F.

func M256MaskMaxPd 

func M256MaskMaxPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm256_mask_max_pd'. Requires AVX512F.

func M256MaskMaxPs 

func M256MaskMaxPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm256_mask_max_ps'. Requires AVX512F.

func M256MaskMinEpi32 

func M256MaskMinEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMinEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINSD'. Intrinsic: '_mm256_mask_min_epi32'. Requires AVX512F.

func M256MaskMinEpi64 

func M256MaskMinEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm256_mask_min_epi64'. Requires AVX512F.

func M256MaskMinEpu32 

func M256MaskMinEpu32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMinEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINUD'. Intrinsic: '_mm256_mask_min_epu32'. Requires AVX512F.

func M256MaskMinEpu64 

func M256MaskMinEpu64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm256_mask_min_epu64'. Requires AVX512F.

func M256MaskMinPd 

func M256MaskMinPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm256_mask_min_pd'. Requires AVX512F.

func M256MaskMinPs 

func M256MaskMinPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm256_mask_min_ps'. Requires AVX512F.

func M256MaskMovEpi32 

func M256MaskMovEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskMovEpi32: Move packed 32-bit integers from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDQA32'. Intrinsic: '_mm256_mask_mov_epi32'. Requires AVX512F.

func M256MaskMovEpi64 

func M256MaskMovEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskMovEpi64: Move packed 64-bit integers from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDQA64'. Intrinsic: '_mm256_mask_mov_epi64'. Requires AVX512F.

func M256MaskMovPd 

func M256MaskMovPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskMovPd: Move packed double-precision (64-bit) floating-point elements from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVAPD'. Intrinsic: '_mm256_mask_mov_pd'. Requires AVX512F.

func M256MaskMovPs 

func M256MaskMovPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskMovPs: Move packed single-precision (32-bit) floating-point elements from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVAPS'. Intrinsic: '_mm256_mask_mov_ps'. Requires AVX512F.

func M256MaskMovedupPd 

func M256MaskMovedupPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
tmp[191:128] := a[191:128]
tmp[255:192] := a[191:128]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm256_mask_movedup_pd'. Requires AVX512F.

func M256MaskMovehdupPs 

func M256MaskMovehdupPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
tmp[159:128] := a[191:160]
tmp[191:160] := a[191:160]
tmp[223:192] := a[255:224]
tmp[255:224] := a[255:224]
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm256_mask_movehdup_ps'. Requires AVX512F.

func M256MaskMoveldupPs 

func M256MaskMoveldupPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
tmp[159:128] := a[159:128]
tmp[191:160] := a[159:128]
tmp[223:192] := a[223:192]
tmp[255:224] := a[223:192]
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm256_mask_moveldup_ps'. Requires AVX512F.

func M256MaskMulEpi32 

func M256MaskMulEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm256_mask_mul_epi32'. Requires AVX512F.

func M256MaskMulEpu32 

func M256MaskMulEpu32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm256_mask_mul_epu32'. Requires AVX512F.

func M256MaskMulPd 

func M256MaskMulPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskMulPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm256_mask_mul_pd'. Requires AVX512F.

func M256MaskMulPs 

func M256MaskMulPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskMulPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). RM. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm256_mask_mul_ps'. Requires AVX512F.

func M256MaskMulloEpi32 

func M256MaskMulloEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskMulloEpi32: Multiply the packed 32-bit integers in 'a' and 'b', producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		tmp[63:0] := a[i+31:i] * b[i+31:i]
		dst[i+31:i] := tmp[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULLD'. Intrinsic: '_mm256_mask_mullo_epi32'. Requires AVX512F.

func M256MaskOrEpi32 

func M256MaskOrEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskOrEpi32: Compute the bitwise OR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] OR b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPORD'. Intrinsic: '_mm256_mask_or_epi32'. Requires AVX512F.

func M256MaskOrEpi64 

func M256MaskOrEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskOrEpi64: Compute the bitwise OR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] OR b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPORQ'. Intrinsic: '_mm256_mask_or_epi64'. Requires AVX512F.

func M256MaskPermutePd 

func M256MaskPermutePd(src x86.M256d, k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
IF (imm8[2] == 0) tmp_dst[191:128] := a[191:128]
IF (imm8[2] == 1) tmp_dst[191:128] := a[255:192]
IF (imm8[3] == 0) tmp_dst[255:192] := a[191:128]
IF (imm8[3] == 1) tmp_dst[255:192] := a[255:192]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm256_mask_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskPermutePs 

func M256MaskPermutePs(src x86.M256, k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M256)

M256MaskPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm256_mask_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskPermutevarPd 

func M256MaskPermutevarPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256i) (dst x86.M256d)

M256MaskPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
IF (b[129] == 0) tmp_dst[191:128] := a[191:128]
IF (b[129] == 1) tmp_dst[191:128] := a[255:192]
IF (b[193] == 0) tmp_dst[255:192] := a[191:128]
IF (b[193] == 1) tmp_dst[255:192] := a[255:192]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm256_mask_permutevar_pd'. Requires AVX512F.

func M256MaskPermutevarPs 

func M256MaskPermutevarPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256i) (dst x86.M256)

M256MaskPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
tmp_dst[159:128] := SELECT4(a[255:128], b[129:128])
tmp_dst[191:160] := SELECT4(a[255:128], b[161:160])
tmp_dst[223:192] := SELECT4(a[255:128], b[193:192])
tmp_dst[255:224] := SELECT4(a[255:128], b[225:224])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm256_mask_permutevar_ps'. Requires AVX512F.

func M256MaskPermutex2varEpi32 

func M256MaskPermutex2varEpi32(a x86.M256i, k x86.Mmask8, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMT2D'. Intrinsic: '_mm256_mask_permutex2var_epi32'. Requires AVX512F.

func M256MaskPermutex2varEpi64 

func M256MaskPermutex2varEpi64(a x86.M256i, k x86.Mmask8, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMT2Q'. Intrinsic: '_mm256_mask_permutex2var_epi64'. Requires AVX512F.

func M256MaskPermutex2varPd 

func M256MaskPermutex2varPd(a x86.M256d, k x86.Mmask8, idx x86.M256i, b x86.M256d) (dst x86.M256d)

M256MaskPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMT2PD'. Intrinsic: '_mm256_mask_permutex2var_pd'. Requires AVX512F.

func M256MaskPermutex2varPs 

func M256MaskPermutex2varPs(a x86.M256, k x86.Mmask8, idx x86.M256i, b x86.M256) (dst x86.M256)

M256MaskPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMT2PS'. Intrinsic: '_mm256_mask_permutex2var_ps'. Requires AVX512F.

func M256MaskPermutexEpi64 

func M256MaskPermutexEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskPermutexEpi64: Shuffle 64-bit integers in 'a' across lanes lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_mask_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskPermutexPd 

func M256MaskPermutexPd(src x86.M256d, k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskPermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_mask_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskPermutexvarEpi32 

func M256MaskPermutexvarEpi32(src x86.M256i, k x86.Mmask8, idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256MaskPermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm256_mask_permutexvar_epi32'. Requires AVX512F.

func M256MaskPermutexvarEpi64 

func M256MaskPermutexvarEpi64(src x86.M256i, k x86.Mmask8, idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256MaskPermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_mask_permutexvar_epi64'. Requires AVX512F.

func M256MaskPermutexvarPd 

func M256MaskPermutexvarPd(src x86.M256d, k x86.Mmask8, idx x86.M256i, a x86.M256d) (dst x86.M256d)

M256MaskPermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_mask_permutexvar_pd'. Requires AVX512F.

func M256MaskPermutexvarPs 

func M256MaskPermutexvarPs(src x86.M256, k x86.Mmask8, idx x86.M256i, a x86.M256) (dst x86.M256)

M256MaskPermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm256_mask_permutexvar_ps'. Requires AVX512F.

func M256MaskRcp14Pd 

func M256MaskRcp14Pd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm256_mask_rcp14_pd'. Requires AVX512F.

func M256MaskRcp14Ps 

func M256MaskRcp14Ps(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm256_mask_rcp14_ps'. Requires AVX512F.

func M256MaskRolEpi32 

func M256MaskRolEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm256_mask_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRolEpi64 

func M256MaskRolEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm256_mask_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRolvEpi32 

func M256MaskRolvEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm256_mask_rolv_epi32'. Requires AVX512F.

func M256MaskRolvEpi64 

func M256MaskRolvEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm256_mask_rolv_epi64'. Requires AVX512F.

func M256MaskRorEpi32 

func M256MaskRorEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm256_mask_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRorEpi64 

func M256MaskRorEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm256_mask_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRorvEpi32 

func M256MaskRorvEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm256_mask_rorv_epi32'. Requires AVX512F.

func M256MaskRorvEpi64 

func M256MaskRorvEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm256_mask_rorv_epi64'. Requires AVX512F.

func M256MaskRoundscalePd 

func M256MaskRoundscalePd(src x86.M256d, k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm256_mask_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRoundscalePs 

func M256MaskRoundscalePs(src x86.M256, k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M256)

M256MaskRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm256_mask_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskRsqrt14Pd 

func M256MaskRsqrt14Pd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm256_mask_rsqrt14_pd'. Requires AVX512F.

func M256MaskRsqrt14Ps 

func M256MaskRsqrt14Ps(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm256_mask_rsqrt14_ps'. Requires AVX512F.

func M256MaskScalefPd 

func M256MaskScalefPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm256_mask_scalef_pd'. Requires AVX512F.

func M256MaskScalefPs 

func M256MaskScalefPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm256_mask_scalef_ps'. Requires AVX512F.

func M256MaskSet1Epi32 

func M256MaskSet1Epi32(src x86.M256i, k x86.Mmask8, a int) (dst x86.M256i)

M256MaskSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm256_mask_set1_epi32'. Requires AVX512F.

func M256MaskSet1Epi64 

func M256MaskSet1Epi64(src x86.M256i, k x86.Mmask8, a int64) (dst x86.M256i)

M256MaskSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm256_mask_set1_epi64'. Requires AVX512F.

func M256MaskShuffleEpi32 

func M256MaskShuffleEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskShuffleEpi32: Shuffle 32-bit integers in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSHUFD'. Intrinsic: '_mm256_mask_shuffle_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShuffleF32x4 

func M256MaskShuffleF32x4(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256, imm8 byte) (dst x86.M256)

M256MaskShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[1])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm256_mask_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShuffleF64x2 

func M256MaskShuffleF64x2(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[1])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm256_mask_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShuffleI32x4 

func M256MaskShuffleI32x4(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[1:0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[3:2])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm256_mask_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShuffleI64x2 

func M256MaskShuffleI64x2(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[1:0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[3:2])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm256_mask_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShufflePd 

func M256MaskShufflePd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskShufflePd: Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]
tmp_dst[191:128] := (imm8[2] == 0) ? a[191:128] : a[255:192]
tmp_dst[255:192] := (imm8[3] == 0) ? b[191:128] : b[255:192]

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm256_mask_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskShufflePs 

func M256MaskShufflePs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256, imm8 byte) (dst x86.M256)

M256MaskShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(b[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(b[255:128], imm8[7:6])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm256_mask_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSllEpi32 

func M256MaskSllEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm256_mask_sll_epi32'. Requires AVX512F.

func M256MaskSllEpi64 

func M256MaskSllEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm256_mask_sll_epi64'. Requires AVX512F.

func M256MaskSlliEpi32 

func M256MaskSlliEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSlliEpi32: Shift packed 32-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm256_mask_slli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSlliEpi64 

func M256MaskSlliEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm256_mask_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSllvEpi32 

func M256MaskSllvEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSllvEpi32: Shift packed 32-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLVD'. Intrinsic: '_mm256_mask_sllv_epi32'. Requires AVX512F.

func M256MaskSllvEpi64 

func M256MaskSllvEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm256_mask_sllv_epi64'. Requires AVX512F.

func M256MaskSqrtPd 

func M256MaskSqrtPd(src x86.M256d, k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm256_mask_sqrt_pd'. Requires AVX512F.

func M256MaskSqrtPs 

func M256MaskSqrtPs(src x86.M256, k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm256_mask_sqrt_ps'. Requires AVX512F.

func M256MaskSraEpi32 

func M256MaskSraEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm256_mask_sra_epi32'. Requires AVX512F.

func M256MaskSraEpi64 

func M256MaskSraEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_mask_sra_epi64'. Requires AVX512F.

func M256MaskSraiEpi32 

func M256MaskSraiEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSraiEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm256_mask_srai_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSraiEpi64 

func M256MaskSraiEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_mask_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSravEpi32 

func M256MaskSravEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSravEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAVD'. Intrinsic: '_mm256_mask_srav_epi32'. Requires AVX512F.

func M256MaskSravEpi64 

func M256MaskSravEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm256_mask_srav_epi64'. Requires AVX512F.

func M256MaskSrlEpi32 

func M256MaskSrlEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm256_mask_srl_epi32'. Requires AVX512F.

func M256MaskSrlEpi64 

func M256MaskSrlEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm256_mask_srl_epi64'. Requires AVX512F.

func M256MaskSrliEpi32 

func M256MaskSrliEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSrliEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm256_mask_srli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSrliEpi64 

func M256MaskSrliEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm256_mask_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskSrlvEpi32 

func M256MaskSrlvEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSrlvEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLVD'. Intrinsic: '_mm256_mask_srlv_epi32'. Requires AVX512F.

func M256MaskSrlvEpi64 

func M256MaskSrlvEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm256_mask_srlv_epi64'. Requires AVX512F.

func M256MaskSubEpi32 

func M256MaskSubEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskSubEpi32: Subtract packed 32-bit integers in 'b' from packed 32-bit integers in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSUBD'. Intrinsic: '_mm256_mask_sub_epi32'. Requires AVX512F.

func M256MaskSubEpi64 

func M256MaskSubEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm256_mask_sub_epi64'. Requires AVX512F.

func M256MaskSubPd 

func M256MaskSubPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskSubPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm256_mask_sub_pd'. Requires AVX512F.

func M256MaskSubPs 

func M256MaskSubPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskSubPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm256_mask_sub_ps'. Requires AVX512F.

func M256MaskTernarylogicEpi32 

func M256MaskTernarylogicEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 32-bit granularity (32-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm256_mask_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskTernarylogicEpi64 

func M256MaskTernarylogicEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 64-bit granularity (64-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm256_mask_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskTestEpi32Mask 

func M256MaskTestEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskTestEpi32Mask: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is non-zero. 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ((a[i+31:i] AND b[i+31:i]) != 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTMD'. Intrinsic: '_mm256_mask_test_epi32_mask'. Requires AVX512F.

func M256MaskTestEpi64Mask 

func M256MaskTestEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskTestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is non-zero. 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm256_mask_test_epi64_mask'. Requires AVX512F.

func M256MaskTestnEpi32Mask 

func M256MaskTestnEpi32Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskTestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 7
	i := j*32
	IF k1[j]
		k[j] := ((a[i+31:i] AND b[i+31:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm256_mask_testn_epi32_mask'. Requires AVX512F.

func M256MaskTestnEpi64Mask 

func M256MaskTestnEpi64Mask(k1 x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256MaskTestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 3
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm256_mask_testn_epi64_mask'. Requires AVX512F.

func M256MaskUnpackhiEpi32 

func M256MaskUnpackhiEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm256_mask_unpackhi_epi32'. Requires AVX512F.

func M256MaskUnpackhiEpi64 

func M256MaskUnpackhiEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm256_mask_unpackhi_epi64'. Requires AVX512F.

func M256MaskUnpackhiPd 

func M256MaskUnpackhiPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm256_mask_unpackhi_pd'. Requires AVX512F.

func M256MaskUnpackhiPs 

func M256MaskUnpackhiPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm256_mask_unpackhi_ps'. Requires AVX512F.

func M256MaskUnpackloEpi32 

func M256MaskUnpackloEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm256_mask_unpacklo_epi32'. Requires AVX512F.

func M256MaskUnpackloEpi64 

func M256MaskUnpackloEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm256_mask_unpacklo_epi64'. Requires AVX512F.

func M256MaskUnpackloPd 

func M256MaskUnpackloPd(src x86.M256d, k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm256_mask_unpacklo_pd'. Requires AVX512F.

func M256MaskUnpackloPs 

func M256MaskUnpackloPs(src x86.M256, k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm256_mask_unpacklo_ps'. Requires AVX512F.

func M256MaskXorEpi32 

func M256MaskXorEpi32(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskXorEpi32: Compute the bitwise XOR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] XOR b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPXORD'. Intrinsic: '_mm256_mask_xor_epi32'. Requires AVX512F.

func M256MaskXorEpi64 

func M256MaskXorEpi64(src x86.M256i, k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskXorEpi64: Compute the bitwise XOR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] XOR b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm256_mask_xor_epi64'. Requires AVX512F.

func M256MaskzAbsEpi32 

func M256MaskzAbsEpi32(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm256_maskz_abs_epi32'. Requires AVX512F.

func M256MaskzAbsEpi64 

func M256MaskzAbsEpi64(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm256_maskz_abs_epi64'. Requires AVX512F.

func M256MaskzAddEpi32 

func M256MaskzAddEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAddEpi32: Add packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPADDD'. Intrinsic: '_mm256_maskz_add_epi32'. Requires AVX512F.

func M256MaskzAddEpi64 

func M256MaskzAddEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] :=0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm256_maskz_add_epi64'. Requires AVX512F.

func M256MaskzAndEpi32 

func M256MaskzAndEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAndEpi32: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] BITWISE AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDD'. Intrinsic: '_mm256_maskz_and_epi32'. Requires AVX512F.

func M256MaskzAndEpi64 

func M256MaskzAndEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAndEpi64: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDQ'. Intrinsic: '_mm256_maskz_and_epi64'. Requires AVX512F.

func M256MaskzAndnotEpi32 

func M256MaskzAndnotEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAndnotEpi32: Compute the bitwise AND NOT of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (NOT a[i+31:i]) AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDND'. Intrinsic: '_mm256_maskz_andnot_epi32'. Requires AVX512F.

func M256MaskzAndnotEpi64 

func M256MaskzAndnotEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzAndnotEpi64: Compute the bitwise AND NOT of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (NOT a[i+63:i]) AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPANDNQ'. Intrinsic: '_mm256_maskz_andnot_epi64'. Requires AVX512F.

func M256MaskzBroadcastF32x4 

func M256MaskzBroadcastF32x4(k x86.Mmask8, a x86.M128) (dst x86.M256)

M256MaskzBroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm256_maskz_broadcast_f32x4'. Requires AVX512F.

func M256MaskzBroadcastI32x4 

func M256MaskzBroadcastI32x4(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzBroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm256_maskz_broadcast_i32x4'. Requires AVX512F.

func M256MaskzBroadcastdEpi32 

func M256MaskzBroadcastdEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm256_maskz_broadcastd_epi32'. Requires AVX512F.

func M256MaskzBroadcastqEpi64 

func M256MaskzBroadcastqEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm256_maskz_broadcastq_epi64'. Requires AVX512F.

func M256MaskzBroadcastsdPd 

func M256MaskzBroadcastsdPd(k x86.Mmask8, a x86.M128d) (dst x86.M256d)

M256MaskzBroadcastsdPd: Broadcast the low double-precision (64-bit) floating-point element from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTSD'. Intrinsic: '_mm256_maskz_broadcastsd_pd'. Requires AVX512F.

func M256MaskzBroadcastssPs 

func M256MaskzBroadcastssPs(k x86.Mmask8, a x86.M128) (dst x86.M256)

M256MaskzBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm256_maskz_broadcastss_ps'. Requires AVX512F.

func M256MaskzCompressEpi32 

func M256MaskzCompressEpi32(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := 0
dst[MAX:256] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm256_maskz_compress_epi32'. Requires AVX512F.

func M256MaskzCompressEpi64 

func M256MaskzCompressEpi64(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := 0
dst[MAX:256] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm256_maskz_compress_epi64'. Requires AVX512F.

func M256MaskzCompressPd 

func M256MaskzCompressPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := 0
dst[MAX:256] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm256_maskz_compress_pd'. Requires AVX512F.

func M256MaskzCompressPs 

func M256MaskzCompressPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[255:m] := 0
dst[MAX:256] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm256_maskz_compress_ps'. Requires AVX512F.

func M256MaskzCvtRoundpsPh 

func M256MaskzCvtRoundpsPh(k x86.Mmask8, a x86.M256, rounding int) (dst x86.M128i)

M256MaskzCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm256_maskz_cvt_roundps_ph'. Requires AVX512F.

func M256MaskzCvtepi16Epi32 

func M256MaskzCvtepi16Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm256_maskz_cvtepi16_epi32'. Requires AVX512F.

func M256MaskzCvtepi16Epi64 

func M256MaskzCvtepi16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepi16Epi64: Sign extend packed 16-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm256_maskz_cvtepi16_epi64'. Requires AVX512F.

func M256MaskzCvtepi32Epi16 

func M256MaskzCvtepi32Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm256_maskz_cvtepi32_epi16'. Requires AVX512F.

func M256MaskzCvtepi32Epi64 

func M256MaskzCvtepi32Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm256_maskz_cvtepi32_epi64'. Requires AVX512F.

func M256MaskzCvtepi32Epi8 

func M256MaskzCvtepi32Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm256_maskz_cvtepi32_epi8'. Requires AVX512F.

func M256MaskzCvtepi32Pd 

func M256MaskzCvtepi32Pd(k x86.Mmask8, a x86.M128i) (dst x86.M256d)

M256MaskzCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm256_maskz_cvtepi32_pd'. Requires AVX512F.

func M256MaskzCvtepi32Ps 

func M256MaskzCvtepi32Ps(k x86.Mmask8, a x86.M256i) (dst x86.M256)

M256MaskzCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm256_maskz_cvtepi32_ps'. Requires AVX512F.

func M256MaskzCvtepi64Epi16 

func M256MaskzCvtepi64Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm256_maskz_cvtepi64_epi16'. Requires AVX512F.

func M256MaskzCvtepi64Epi32 

func M256MaskzCvtepi64Epi32(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm256_maskz_cvtepi64_epi32'. Requires AVX512F.

func M256MaskzCvtepi64Epi8 

func M256MaskzCvtepi64Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm256_maskz_cvtepi64_epi8'. Requires AVX512F.

func M256MaskzCvtepi8Epi32 

func M256MaskzCvtepi8Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepi8Epi32: Sign extend packed 8-bit integers in the low 8 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm256_maskz_cvtepi8_epi32'. Requires AVX512F.

func M256MaskzCvtepi8Epi64 

func M256MaskzCvtepi8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepi8Epi64: Sign extend packed 8-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm256_maskz_cvtepi8_epi64'. Requires AVX512F.

func M256MaskzCvtepu16Epi32 

func M256MaskzCvtepu16Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm256_maskz_cvtepu16_epi32'. Requires AVX512F.

func M256MaskzCvtepu16Epi64 

func M256MaskzCvtepu16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm256_maskz_cvtepu16_epi64'. Requires AVX512F.

func M256MaskzCvtepu32Epi64 

func M256MaskzCvtepu32Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm256_maskz_cvtepu32_epi64'. Requires AVX512F.

func M256MaskzCvtepu32Pd 

func M256MaskzCvtepu32Pd(k x86.Mmask8, a x86.M128i) (dst x86.M256d)

M256MaskzCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm256_maskz_cvtepu32_pd'. Requires AVX512F.

func M256MaskzCvtepu8Epi32 

func M256MaskzCvtepu8Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in the low 8 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm256_maskz_cvtepu8_epi32'. Requires AVX512F.

func M256MaskzCvtepu8Epi64 

func M256MaskzCvtepu8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M256i)

M256MaskzCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm256_maskz_cvtepu8_epi64'. Requires AVX512F.

func M256MaskzCvtpdEpi32 

func M256MaskzCvtpdEpi32(k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskzCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm256_maskz_cvtpd_epi32'. Requires AVX512F.

func M256MaskzCvtpdEpu32 

func M256MaskzCvtpdEpu32(k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskzCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm256_maskz_cvtpd_epu32'. Requires AVX512F.

func M256MaskzCvtpdPs 

func M256MaskzCvtpdPs(k x86.Mmask8, a x86.M256d) (dst x86.M128)

M256MaskzCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm256_maskz_cvtpd_ps'. Requires AVX512F.

func M256MaskzCvtphPs 

func M256MaskzCvtphPs(k x86.Mmask8, a x86.M128i) (dst x86.M256)

M256MaskzCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm256_maskz_cvtph_ps'. Requires AVX512F.

func M256MaskzCvtpsEpi32 

func M256MaskzCvtpsEpi32(k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskzCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm256_maskz_cvtps_epi32'. Requires AVX512F.

func M256MaskzCvtpsEpu32 

func M256MaskzCvtpsEpu32(k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskzCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm256_maskz_cvtps_epu32'. Requires AVX512F.

func M256MaskzCvtpsPh 

func M256MaskzCvtpsPh(k x86.Mmask8, a x86.M256, rounding int) (dst x86.M128i)

M256MaskzCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm256_maskz_cvtps_ph'. Requires AVX512F.

func M256MaskzCvtsepi32Epi16 

func M256MaskzCvtsepi32Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm256_maskz_cvtsepi32_epi16'. Requires AVX512F.

func M256MaskzCvtsepi32Epi8 

func M256MaskzCvtsepi32Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm256_maskz_cvtsepi32_epi8'. Requires AVX512F.

func M256MaskzCvtsepi64Epi16 

func M256MaskzCvtsepi64Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm256_maskz_cvtsepi64_epi16'. Requires AVX512F.

func M256MaskzCvtsepi64Epi32 

func M256MaskzCvtsepi64Epi32(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm256_maskz_cvtsepi64_epi32'. Requires AVX512F.

func M256MaskzCvtsepi64Epi8 

func M256MaskzCvtsepi64Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm256_maskz_cvtsepi64_epi8'. Requires AVX512F.

func M256MaskzCvttpdEpi32 

func M256MaskzCvttpdEpi32(k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskzCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm256_maskz_cvttpd_epi32'. Requires AVX512F.

func M256MaskzCvttpdEpu32 

func M256MaskzCvttpdEpu32(k x86.Mmask8, a x86.M256d) (dst x86.M128i)

M256MaskzCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm256_maskz_cvttpd_epu32'. Requires AVX512F.

func M256MaskzCvttpsEpi32 

func M256MaskzCvttpsEpi32(k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskzCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*i
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_IntegerTruncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm256_maskz_cvttps_epi32'. Requires AVX512F.

func M256MaskzCvttpsEpu32 

func M256MaskzCvttpsEpu32(k x86.Mmask8, a x86.M256) (dst x86.M256i)

M256MaskzCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm256_maskz_cvttps_epu32'. Requires AVX512F.

func M256MaskzCvtusepi32Epi16 

func M256MaskzCvtusepi32Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm256_maskz_cvtusepi32_epi16'. Requires AVX512F.

func M256MaskzCvtusepi32Epi8 

func M256MaskzCvtusepi32Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm256_maskz_cvtusepi32_epi8'. Requires AVX512F.

func M256MaskzCvtusepi64Epi16 

func M256MaskzCvtusepi64Epi16(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm256_maskz_cvtusepi64_epi16'. Requires AVX512F.

func M256MaskzCvtusepi64Epi32 

func M256MaskzCvtusepi64Epi32(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm256_maskz_cvtusepi64_epi32'. Requires AVX512F.

func M256MaskzCvtusepi64Epi8 

func M256MaskzCvtusepi64Epi8(k x86.Mmask8, a x86.M256i) (dst x86.M128i)

M256MaskzCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm256_maskz_cvtusepi64_epi8'. Requires AVX512F.

func M256MaskzDivPd 

func M256MaskzDivPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm256_maskz_div_pd'. Requires AVX512F.

func M256MaskzDivPs 

func M256MaskzDivPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm256_maskz_div_ps'. Requires AVX512F.

func M256MaskzExpandEpi32 

func M256MaskzExpandEpi32(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm256_maskz_expand_epi32'. Requires AVX512F.

func M256MaskzExpandEpi64 

func M256MaskzExpandEpi64(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm256_maskz_expand_epi64'. Requires AVX512F.

func M256MaskzExpandPd 

func M256MaskzExpandPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm256_maskz_expand_pd'. Requires AVX512F.

func M256MaskzExpandPs 

func M256MaskzExpandPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm256_maskz_expand_ps'. Requires AVX512F.

func M256MaskzExtractf32x4Ps 

func M256MaskzExtractf32x4Ps(k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M128)

M256MaskzExtractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[127:0] := a[127:0]
1: tmp[127:0] := a[255:128]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm256_maskz_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzExtracti32x4Epi32 

func M256MaskzExtracti32x4Epi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M128i)

M256MaskzExtracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm256_maskz_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzFixupimmPd 

func M256MaskzFixupimmPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256i, imm8 byte) (dst x86.M256d)

M256MaskzFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm256_maskz_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzFixupimmPs 

func M256MaskzFixupimmPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256i, imm8 byte) (dst x86.M256)

M256MaskzFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm256_maskz_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzFmaddPd 

func M256MaskzFmaddPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm256_maskz_fmadd_pd'. Requires AVX512F.

func M256MaskzFmaddPs 

func M256MaskzFmaddPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm256_maskz_fmadd_ps'. Requires AVX512F.

func M256MaskzFmaddsubPd 

func M256MaskzFmaddsubPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm256_maskz_fmaddsub_pd'. Requires AVX512F.

func M256MaskzFmaddsubPs 

func M256MaskzFmaddsubPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm256_maskz_fmaddsub_ps'. Requires AVX512F.

func M256MaskzFmsubPd 

func M256MaskzFmsubPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm256_maskz_fmsub_pd'. Requires AVX512F.

func M256MaskzFmsubPs 

func M256MaskzFmsubPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm256_maskz_fmsub_ps'. Requires AVX512F.

func M256MaskzFmsubaddPd 

func M256MaskzFmsubaddPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm256_maskz_fmsubadd_pd'. Requires AVX512F.

func M256MaskzFmsubaddPs 

func M256MaskzFmsubaddPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm256_maskz_fmsubadd_ps'. Requires AVX512F.

func M256MaskzFnmaddPd 

func M256MaskzFnmaddPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm256_maskz_fnmadd_pd'. Requires AVX512F.

func M256MaskzFnmaddPs 

func M256MaskzFnmaddPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm256_maskz_fnmadd_ps'. Requires AVX512F.

func M256MaskzFnmsubPd 

func M256MaskzFnmsubPd(k x86.Mmask8, a x86.M256d, b x86.M256d, c x86.M256d) (dst x86.M256d)

M256MaskzFnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm256_maskz_fnmsub_pd'. Requires AVX512F.

func M256MaskzFnmsubPs 

func M256MaskzFnmsubPs(k x86.Mmask8, a x86.M256, b x86.M256, c x86.M256) (dst x86.M256)

M256MaskzFnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm256_maskz_fnmsub_ps'. Requires AVX512F.

func M256MaskzGetexpPd 

func M256MaskzGetexpPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzGetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm256_maskz_getexp_pd'. Requires AVX512F.

func M256MaskzGetexpPs 

func M256MaskzGetexpPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzGetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm256_maskz_getexp_ps'. Requires AVX512F.

func M256MaskzGetmantPd 

func M256MaskzGetmantPd(k x86.Mmask8, a x86.M256d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256d)

M256MaskzGetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm256_maskz_getmant_pd'. Requires AVX512F.

func M256MaskzGetmantPs 

func M256MaskzGetmantPs(k x86.Mmask8, a x86.M256, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M256)

M256MaskzGetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 7
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm256_maskz_getmant_ps'. Requires AVX512F.

func M256MaskzInsertf32x4 

func M256MaskzInsertf32x4(k x86.Mmask8, a x86.M256, b x86.M128, imm8 byte) (dst x86.M256)

M256MaskzInsertf32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
ESAC
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm256_maskz_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzInserti32x4 

func M256MaskzInserti32x4(k x86.Mmask8, a x86.M256i, b x86.M128i, imm8 byte) (dst x86.M256i)

M256MaskzInserti32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[255:0] := a[255:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
ESAC
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm256_maskz_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzMaxEpi32 

func M256MaskzMaxEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMaxEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXSD'. Intrinsic: '_mm256_maskz_max_epi32'. Requires AVX512F.

func M256MaskzMaxEpi64 

func M256MaskzMaxEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm256_maskz_max_epi64'. Requires AVX512F.

func M256MaskzMaxEpu32 

func M256MaskzMaxEpu32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMaxEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXUD'. Intrinsic: '_mm256_maskz_max_epu32'. Requires AVX512F.

func M256MaskzMaxEpu64 

func M256MaskzMaxEpu64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm256_maskz_max_epu64'. Requires AVX512F.

func M256MaskzMaxPd 

func M256MaskzMaxPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm256_maskz_max_pd'. Requires AVX512F.

func M256MaskzMaxPs 

func M256MaskzMaxPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm256_maskz_max_ps'. Requires AVX512F.

func M256MaskzMinEpi32 

func M256MaskzMinEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMinEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINSD'. Intrinsic: '_mm256_maskz_min_epi32'. Requires AVX512F.

func M256MaskzMinEpi64 

func M256MaskzMinEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm256_maskz_min_epi64'. Requires AVX512F.

func M256MaskzMinEpu32 

func M256MaskzMinEpu32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMinEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINUD'. Intrinsic: '_mm256_maskz_min_epu32'. Requires AVX512F.

func M256MaskzMinEpu64 

func M256MaskzMinEpu64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm256_maskz_min_epu64'. Requires AVX512F.

func M256MaskzMinPd 

func M256MaskzMinPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm256_maskz_min_pd'. Requires AVX512F.

func M256MaskzMinPs 

func M256MaskzMinPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm256_maskz_min_ps'. Requires AVX512F.

func M256MaskzMovEpi32 

func M256MaskzMovEpi32(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzMovEpi32: Move packed 32-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDQA32'. Intrinsic: '_mm256_maskz_mov_epi32'. Requires AVX512F.

func M256MaskzMovEpi64 

func M256MaskzMovEpi64(k x86.Mmask8, a x86.M256i) (dst x86.M256i)

M256MaskzMovEpi64: Move packed 64-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDQA64'. Intrinsic: '_mm256_maskz_mov_epi64'. Requires AVX512F.

func M256MaskzMovPd 

func M256MaskzMovPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzMovPd: Move packed double-precision (64-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVAPD'. Intrinsic: '_mm256_maskz_mov_pd'. Requires AVX512F.

func M256MaskzMovPs 

func M256MaskzMovPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzMovPs: Move packed single-precision (32-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVAPS'. Intrinsic: '_mm256_maskz_mov_ps'. Requires AVX512F.

func M256MaskzMovedupPd 

func M256MaskzMovedupPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
tmp[191:128] := a[191:128]
tmp[255:192] := a[191:128]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm256_maskz_movedup_pd'. Requires AVX512F.

func M256MaskzMovehdupPs 

func M256MaskzMovehdupPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
tmp[159:128] := a[191:160]
tmp[191:160] := a[191:160]
tmp[223:192] := a[255:224]
tmp[255:224] := a[255:224]
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm256_maskz_movehdup_ps'. Requires AVX512F.

func M256MaskzMoveldupPs 

func M256MaskzMoveldupPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
tmp[159:128] := a[159:128]
tmp[191:160] := a[159:128]
tmp[223:192] := a[223:192]
tmp[255:224] := a[223:192]
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm256_maskz_moveldup_ps'. Requires AVX512F.

func M256MaskzMulEpi32 

func M256MaskzMulEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm256_maskz_mul_epi32'. Requires AVX512F.

func M256MaskzMulEpu32 

func M256MaskzMulEpu32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm256_maskz_mul_epu32'. Requires AVX512F.

func M256MaskzMulPd 

func M256MaskzMulPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzMulPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm256_maskz_mul_pd'. Requires AVX512F.

func M256MaskzMulPs 

func M256MaskzMulPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzMulPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm256_maskz_mul_ps'. Requires AVX512F.

func M256MaskzMulloEpi32 

func M256MaskzMulloEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzMulloEpi32: Multiply the packed 32-bit integers in 'a' and 'b', producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		tmp[63:0] := a[i+31:i] * b[i+31:i]
		dst[i+31:i] := tmp[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMULLD'. Intrinsic: '_mm256_maskz_mullo_epi32'. Requires AVX512F.

func M256MaskzOrEpi32 

func M256MaskzOrEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzOrEpi32: Compute the bitwise OR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] OR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPORD'. Intrinsic: '_mm256_maskz_or_epi32'. Requires AVX512F.

func M256MaskzOrEpi64 

func M256MaskzOrEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzOrEpi64: Compute the bitwise OR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] OR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPORQ'. Intrinsic: '_mm256_maskz_or_epi64'. Requires AVX512F.

func M256MaskzPermutePd 

func M256MaskzPermutePd(k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskzPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
IF (imm8[2] == 0) tmp_dst[191:128] := a[191:128]
IF (imm8[2] == 1) tmp_dst[191:128] := a[255:192]
IF (imm8[3] == 0) tmp_dst[255:192] := a[191:128]
IF (imm8[3] == 1) tmp_dst[255:192] := a[255:192]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm256_maskz_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzPermutePs 

func M256MaskzPermutePs(k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M256)

M256MaskzPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm256_maskz_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzPermutevarPd 

func M256MaskzPermutevarPd(k x86.Mmask8, a x86.M256d, b x86.M256i) (dst x86.M256d)

M256MaskzPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
IF (b[129] == 0) tmp_dst[191:128] := a[191:128]
IF (b[129] == 1) tmp_dst[191:128] := a[255:192]
IF (b[193] == 0) tmp_dst[255:192] := a[191:128]
IF (b[193] == 1) tmp_dst[255:192] := a[255:192]
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm256_maskz_permutevar_pd'. Requires AVX512F.

func M256MaskzPermutevarPs 

func M256MaskzPermutevarPs(k x86.Mmask8, a x86.M256, b x86.M256i) (dst x86.M256)

M256MaskzPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
tmp_dst[159:128] := SELECT4(a[255:128], b[129:128])
tmp_dst[191:160] := SELECT4(a[255:128], b[161:160])
tmp_dst[223:192] := SELECT4(a[255:128], b[193:192])
tmp_dst[255:224] := SELECT4(a[255:128], b[225:224])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm256_maskz_permutevar_ps'. Requires AVX512F.

func M256MaskzPermutex2varEpi32 

func M256MaskzPermutex2varEpi32(k x86.Mmask8, a x86.M256i, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+3]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm256_maskz_permutex2var_epi32'. Requires AVX512F.

func M256MaskzPermutex2varEpi64 

func M256MaskzPermutex2varEpi64(k x86.Mmask8, a x86.M256i, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+2]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm256_maskz_permutex2var_epi64'. Requires AVX512F.

func M256MaskzPermutex2varPd 

func M256MaskzPermutex2varPd(k x86.Mmask8, a x86.M256d, idx x86.M256i, b x86.M256d) (dst x86.M256d)

M256MaskzPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+2]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm256_maskz_permutex2var_pd'. Requires AVX512F.

func M256MaskzPermutex2varPs 

func M256MaskzPermutex2varPs(k x86.Mmask8, a x86.M256, idx x86.M256i, b x86.M256) (dst x86.M256)

M256MaskzPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+3]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm256_maskz_permutex2var_ps'. Requires AVX512F.

func M256MaskzPermutexEpi64 

func M256MaskzPermutexEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzPermutexEpi64: Shuffle 64-bit integers in 'a' across lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_maskz_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzPermutexPd 

func M256MaskzPermutexPd(k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskzPermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_maskz_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzPermutexvarEpi32 

func M256MaskzPermutexvarEpi32(k x86.Mmask8, idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256MaskzPermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm256_maskz_permutexvar_epi32'. Requires AVX512F.

func M256MaskzPermutexvarEpi64 

func M256MaskzPermutexvarEpi64(k x86.Mmask8, idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256MaskzPermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_maskz_permutexvar_epi64'. Requires AVX512F.

func M256MaskzPermutexvarPd 

func M256MaskzPermutexvarPd(k x86.Mmask8, idx x86.M256i, a x86.M256d) (dst x86.M256d)

M256MaskzPermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_maskz_permutexvar_pd'. Requires AVX512F.

func M256MaskzPermutexvarPs 

func M256MaskzPermutexvarPs(k x86.Mmask8, idx x86.M256i, a x86.M256) (dst x86.M256)

M256MaskzPermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm256_maskz_permutexvar_ps'. Requires AVX512F.

func M256MaskzRcp14Pd 

func M256MaskzRcp14Pd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm256_maskz_rcp14_pd'. Requires AVX512F.

func M256MaskzRcp14Ps 

func M256MaskzRcp14Ps(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm256_maskz_rcp14_ps'. Requires AVX512F.

func M256MaskzRolEpi32 

func M256MaskzRolEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm256_maskz_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRolEpi64 

func M256MaskzRolEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm256_maskz_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRolvEpi32 

func M256MaskzRolvEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm256_maskz_rolv_epi32'. Requires AVX512F.

func M256MaskzRolvEpi64 

func M256MaskzRolvEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm256_maskz_rolv_epi64'. Requires AVX512F.

func M256MaskzRorEpi32 

func M256MaskzRorEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm256_maskz_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRorEpi64 

func M256MaskzRorEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm256_maskz_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRorvEpi32 

func M256MaskzRorvEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm256_maskz_rorv_epi32'. Requires AVX512F.

func M256MaskzRorvEpi64 

func M256MaskzRorvEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm256_maskz_rorv_epi64'. Requires AVX512F.

func M256MaskzRoundscalePd 

func M256MaskzRoundscalePd(k x86.Mmask8, a x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskzRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm256_maskz_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRoundscalePs 

func M256MaskzRoundscalePs(k x86.Mmask8, a x86.M256, imm8 byte) (dst x86.M256)

M256MaskzRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm256_maskz_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzRsqrt14Pd 

func M256MaskzRsqrt14Pd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm256_maskz_rsqrt14_pd'. Requires AVX512F.

func M256MaskzRsqrt14Ps 

func M256MaskzRsqrt14Ps(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm256_maskz_rsqrt14_ps'. Requires AVX512F.

func M256MaskzScalefPd 

func M256MaskzScalefPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm256_maskz_scalef_pd'. Requires AVX512F.

func M256MaskzScalefPs 

func M256MaskzScalefPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm256_maskz_scalef_ps'. Requires AVX512F.

func M256MaskzSet1Epi32 

func M256MaskzSet1Epi32(k x86.Mmask8, a int) (dst x86.M256i)

M256MaskzSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm256_maskz_set1_epi32'. Requires AVX512F.

func M256MaskzSet1Epi64 

func M256MaskzSet1Epi64(k x86.Mmask8, a int64) (dst x86.M256i)

M256MaskzSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm256_maskz_set1_epi64'. Requires AVX512F.

func M256MaskzShuffleEpi32 

func M256MaskzShuffleEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzShuffleEpi32: Shuffle 32-bit integers in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSHUFD'. Intrinsic: '_mm256_maskz_shuffle_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShuffleF32x4 

func M256MaskzShuffleF32x4(k x86.Mmask8, a x86.M256, b x86.M256, imm8 byte) (dst x86.M256)

M256MaskzShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[1])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm256_maskz_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShuffleF64x2 

func M256MaskzShuffleF64x2(k x86.Mmask8, a x86.M256d, b x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskzShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[1])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm256_maskz_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShuffleI32x4 

func M256MaskzShuffleI32x4(k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[1:0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[3:2])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm256_maskz_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShuffleI64x2 

func M256MaskzShuffleI64x2(k x86.Mmask8, a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT2(a[255:0], imm8[1:0])
tmp_dst[255:128] := SELECT2(b[255:0], imm8[3:2])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm256_maskz_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShufflePd 

func M256MaskzShufflePd(k x86.Mmask8, a x86.M256d, b x86.M256d, imm8 byte) (dst x86.M256d)

M256MaskzShufflePd: Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]
tmp_dst[191:128] := (imm8[2] == 0) ? a[191:128] : a[255:192]
tmp_dst[255:192] := (imm8[3] == 0) ? b[191:128] : b[255:192]

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm256_maskz_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzShufflePs 

func M256MaskzShufflePs(k x86.Mmask8, a x86.M256, b x86.M256, imm8 byte) (dst x86.M256)

M256MaskzShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(b[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(b[255:128], imm8[7:6])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm256_maskz_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSllEpi32 

func M256MaskzSllEpi32(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm256_maskz_sll_epi32'. Requires AVX512F.

func M256MaskzSllEpi64 

func M256MaskzSllEpi64(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm256_maskz_sll_epi64'. Requires AVX512F.

func M256MaskzSlliEpi32 

func M256MaskzSlliEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSlliEpi32: Shift packed 32-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm256_maskz_slli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSlliEpi64 

func M256MaskzSlliEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm256_maskz_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSllvEpi32 

func M256MaskzSllvEpi32(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSllvEpi32: Shift packed 32-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLVD'. Intrinsic: '_mm256_maskz_sllv_epi32'. Requires AVX512F.

func M256MaskzSllvEpi64 

func M256MaskzSllvEpi64(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm256_maskz_sllv_epi64'. Requires AVX512F.

func M256MaskzSqrtPd 

func M256MaskzSqrtPd(k x86.Mmask8, a x86.M256d) (dst x86.M256d)

M256MaskzSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm256_maskz_sqrt_pd'. Requires AVX512F.

func M256MaskzSqrtPs 

func M256MaskzSqrtPs(k x86.Mmask8, a x86.M256) (dst x86.M256)

M256MaskzSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm256_maskz_sqrt_ps'. Requires AVX512F.

func M256MaskzSraEpi32 

func M256MaskzSraEpi32(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm256_maskz_sra_epi32'. Requires AVX512F.

func M256MaskzSraEpi64 

func M256MaskzSraEpi64(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_maskz_sra_epi64'. Requires AVX512F.

func M256MaskzSraiEpi32 

func M256MaskzSraiEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSraiEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm256_maskz_srai_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSraiEpi64 

func M256MaskzSraiEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_maskz_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSravEpi32 

func M256MaskzSravEpi32(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSravEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAVD'. Intrinsic: '_mm256_maskz_srav_epi32'. Requires AVX512F.

func M256MaskzSravEpi64 

func M256MaskzSravEpi64(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm256_maskz_srav_epi64'. Requires AVX512F.

func M256MaskzSrlEpi32 

func M256MaskzSrlEpi32(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm256_maskz_srl_epi32'. Requires AVX512F.

func M256MaskzSrlEpi64 

func M256MaskzSrlEpi64(k x86.Mmask8, a x86.M256i, count x86.M128i) (dst x86.M256i)

M256MaskzSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm256_maskz_srl_epi64'. Requires AVX512F.

func M256MaskzSrliEpi32 

func M256MaskzSrliEpi32(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSrliEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm256_maskz_srli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSrliEpi64 

func M256MaskzSrliEpi64(k x86.Mmask8, a x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm256_maskz_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzSrlvEpi32 

func M256MaskzSrlvEpi32(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSrlvEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLVD'. Intrinsic: '_mm256_maskz_srlv_epi32'. Requires AVX512F.

func M256MaskzSrlvEpi64 

func M256MaskzSrlvEpi64(k x86.Mmask8, a x86.M256i, count x86.M256i) (dst x86.M256i)

M256MaskzSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm256_maskz_srlv_epi64'. Requires AVX512F.

func M256MaskzSubEpi32 

func M256MaskzSubEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzSubEpi32: Subtract packed 32-bit integers in 'b' from packed 32-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSUBD'. Intrinsic: '_mm256_maskz_sub_epi32'. Requires AVX512F.

func M256MaskzSubEpi64 

func M256MaskzSubEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm256_maskz_sub_epi64'. Requires AVX512F.

func M256MaskzSubPd 

func M256MaskzSubPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzSubPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm256_maskz_sub_pd'. Requires AVX512F.

func M256MaskzSubPs 

func M256MaskzSubPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzSubPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm256_maskz_sub_ps'. Requires AVX512F.

func M256MaskzTernarylogicEpi32 

func M256MaskzTernarylogicEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i, c x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm256_maskz_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzTernarylogicEpi64 

func M256MaskzTernarylogicEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i, c x86.M256i, imm8 byte) (dst x86.M256i)

M256MaskzTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm256_maskz_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256MaskzUnpackhiEpi32 

func M256MaskzUnpackhiEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm256_maskz_unpackhi_epi32'. Requires AVX512F.

func M256MaskzUnpackhiEpi64 

func M256MaskzUnpackhiEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm256_maskz_unpackhi_epi64'. Requires AVX512F.

func M256MaskzUnpackhiPd 

func M256MaskzUnpackhiPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm256_maskz_unpackhi_pd'. Requires AVX512F.

func M256MaskzUnpackhiPs 

func M256MaskzUnpackhiPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm256_maskz_unpackhi_ps'. Requires AVX512F.

func M256MaskzUnpackloEpi32 

func M256MaskzUnpackloEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm256_maskz_unpacklo_epi32'. Requires AVX512F.

func M256MaskzUnpackloEpi64 

func M256MaskzUnpackloEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm256_maskz_unpacklo_epi64'. Requires AVX512F.

func M256MaskzUnpackloPd 

func M256MaskzUnpackloPd(k x86.Mmask8, a x86.M256d, b x86.M256d) (dst x86.M256d)

M256MaskzUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm256_maskz_unpacklo_pd'. Requires AVX512F.

func M256MaskzUnpackloPs 

func M256MaskzUnpackloPs(k x86.Mmask8, a x86.M256, b x86.M256) (dst x86.M256)

M256MaskzUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm256_maskz_unpacklo_ps'. Requires AVX512F.

func M256MaskzXorEpi32 

func M256MaskzXorEpi32(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzXorEpi32: Compute the bitwise XOR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] XOR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPXORD'. Intrinsic: '_mm256_maskz_xor_epi32'. Requires AVX512F.

func M256MaskzXorEpi64 

func M256MaskzXorEpi64(k x86.Mmask8, a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaskzXorEpi64: Compute the bitwise XOR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] XOR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm256_maskz_xor_epi64'. Requires AVX512F.

func M256MaxEpi64 

func M256MaxEpi64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm256_max_epi64'. Requires AVX512F.

func M256MaxEpu64 

func M256MaxEpu64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm256_max_epu64'. Requires AVX512F.

func M256MinEpi64 

func M256MinEpi64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm256_min_epi64'. Requires AVX512F.

func M256MinEpu64 

func M256MinEpu64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256MinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm256_min_epu64'. Requires AVX512F.

func M256Permutex2varEpi32 

func M256Permutex2varEpi32(a x86.M256i, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm256_permutex2var_epi32'. Requires AVX512F.

func M256Permutex2varEpi64 

func M256Permutex2varEpi64(a x86.M256i, idx x86.M256i, b x86.M256i) (dst x86.M256i)

M256Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm256_permutex2var_epi64'. Requires AVX512F.

func M256Permutex2varPd 

func M256Permutex2varPd(a x86.M256d, idx x86.M256i, b x86.M256d) (dst x86.M256d)

M256Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	off := idx[i+1:i]*64
	dst[i+63:i] := idx[i+2] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm256_permutex2var_pd'. Requires AVX512F.

func M256Permutex2varPs 

func M256Permutex2varPs(a x86.M256, idx x86.M256i, b x86.M256) (dst x86.M256)

M256Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	off := idx[i+2:i]*32
	dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm256_permutex2var_ps'. Requires AVX512F.

func M256PermutexEpi64 

func M256PermutexEpi64(a x86.M256i, imm8 byte) (dst x86.M256i)

M256PermutexEpi64: Shuffle 64-bit integers in 'a' across lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

dst[63:0] := SELECT4(a[255:0], imm8[1:0])
dst[127:64] := SELECT4(a[255:0], imm8[3:2])
dst[191:128] := SELECT4(a[255:0], imm8[5:4])
dst[255:192] := SELECT4(a[255:0], imm8[7:6])
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256PermutexPd 

func M256PermutexPd(a x86.M256d, imm8 byte) (dst x86.M256d)

M256PermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

dst[63:0] := SELECT4(a[255:0], imm8[1:0])
dst[127:64] := SELECT4(a[255:0], imm8[3:2])
dst[191:128] := SELECT4(a[255:0], imm8[5:4])
dst[255:192] := SELECT4(a[255:0], imm8[7:6])
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256PermutexvarEpi32 

func M256PermutexvarEpi32(idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256PermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	dst[i+31:i] := a[id+31:id]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm256_permutexvar_epi32'. Requires AVX512F.

func M256PermutexvarEpi64 

func M256PermutexvarEpi64(idx x86.M256i, a x86.M256i) (dst x86.M256i)

M256PermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	dst[i+63:i] := a[id+63:id]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm256_permutexvar_epi64'. Requires AVX512F.

func M256PermutexvarPd 

func M256PermutexvarPd(idx x86.M256i, a x86.M256d) (dst x86.M256d)

M256PermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	id := idx[i+1:i]*64
	dst[i+63:i] := a[id+63:id]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm256_permutexvar_pd'. Requires AVX512F.

func M256PermutexvarPs 

func M256PermutexvarPs(idx x86.M256i, a x86.M256) (dst x86.M256)

M256PermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx'. 

FOR j := 0 to 7
	i := j*32
	id := idx[i+2:i]*32
	dst[i+31:i] := a[id+31:id]
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm256_permutexvar_ps'. Requires AVX512F.

func M256Rcp14Pd 

func M256Rcp14Pd(a x86.M256d) (dst x86.M256d)

M256Rcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm256_rcp14_pd'. Requires AVX512F.

func M256Rcp14Ps 

func M256Rcp14Ps(a x86.M256) (dst x86.M256)

M256Rcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm256_rcp14_ps'. Requires AVX512F.

func M256RolEpi32 

func M256RolEpi32(a x86.M256i, imm8 byte) (dst x86.M256i)

M256RolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm256_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256RolEpi64 

func M256RolEpi64(a x86.M256i, imm8 byte) (dst x86.M256i)

M256RolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm256_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256RolvEpi32 

func M256RolvEpi32(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256RolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm256_rolv_epi32'. Requires AVX512F.

func M256RolvEpi64 

func M256RolvEpi64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256RolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm256_rolv_epi64'. Requires AVX512F.

func M256RorEpi32 

func M256RorEpi32(a x86.M256i, imm8 byte) (dst x86.M256i)

M256RorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm256_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256RorEpi64 

func M256RorEpi64(a x86.M256i, imm8 byte) (dst x86.M256i)

M256RorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm256_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256RorvEpi32 

func M256RorvEpi32(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256RorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm256_rorv_epi32'. Requires AVX512F.

func M256RorvEpi64 

func M256RorvEpi64(a x86.M256i, b x86.M256i) (dst x86.M256i)

M256RorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm256_rorv_epi64'. Requires AVX512F.

func M256RoundscalePd 

func M256RoundscalePd(a x86.M256d, imm8 byte) (dst x86.M256d)

M256RoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm256_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256RoundscalePs 

func M256RoundscalePs(a x86.M256, imm8 byte) (dst x86.M256)

M256RoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm256_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256ScalefPd 

func M256ScalefPd(a x86.M256d, b x86.M256d) (dst x86.M256d)

M256ScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm256_scalef_pd'. Requires AVX512F.

func M256ScalefPs 

func M256ScalefPs(a x86.M256, b x86.M256) (dst x86.M256)

M256ScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 7
	i := j*32
	dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm256_scalef_ps'. Requires AVX512F.

func M256ShuffleF32x4 

func M256ShuffleF32x4(a x86.M256, b x86.M256, imm8 byte) (dst x86.M256)

M256ShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT2(a[255:0], imm8[0])
dst[255:128] := SELECT2(b[255:0], imm8[1])
dst[MAX:256] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm256_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256ShuffleF64x2 

func M256ShuffleF64x2(a x86.M256d, b x86.M256d, imm8 byte) (dst x86.M256d)

M256ShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT2(a[255:0], imm8[0])
dst[255:128] := SELECT2(b[255:0], imm8[1])
dst[MAX:256] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm256_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256ShuffleI32x4 

func M256ShuffleI32x4(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256ShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT2(a[255:0], imm8[1:0])
dst[255:128] := SELECT2(b[255:0], imm8[3:2])
dst[MAX:256] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm256_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256ShuffleI64x2 

func M256ShuffleI64x2(a x86.M256i, b x86.M256i, imm8 byte) (dst x86.M256i)

M256ShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT2(src, control){
	CASE(control[0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT2(a[255:0], imm8[1:0])
dst[255:128] := SELECT2(b[255:0], imm8[3:2])
dst[MAX:256] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm256_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256SraEpi64 

func M256SraEpi64(a x86.M256i, count x86.M128i) (dst x86.M256i)

M256SraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF count[63:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_sra_epi64'. Requires AVX512F.

func M256SraiEpi64 

func M256SraiEpi64(a x86.M256i, imm8 byte) (dst x86.M256i)

M256SraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	IF imm8[7:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm256_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256SravEpi64 

func M256SravEpi64(a x86.M256i, count x86.M256i) (dst x86.M256i)

M256SravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*64
	dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm256_srav_epi64'. Requires AVX512F.

func M256TernarylogicEpi32 

func M256TernarylogicEpi32(a x86.M256i, b x86.M256i, c x86.M256i, imm8 byte) (dst x86.M256i)

M256TernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 7
	i := j*32
	FOR h := 0 to 31
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm256_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256TernarylogicEpi64 

func M256TernarylogicEpi64(a x86.M256i, b x86.M256i, c x86.M256i, imm8 byte) (dst x86.M256i)

M256TernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 3
	i := j*64
	FOR h := 0 to 63
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm256_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M256TestEpi32Mask 

func M256TestEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256TestEpi32Mask: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is non-zero. 

FOR j := 0 to 7
	i := j*32
	k[j] := ((a[i+31:i] AND b[i+31:i]) != 0) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTMD'. Intrinsic: '_mm256_test_epi32_mask'. Requires AVX512F.

func M256TestEpi64Mask 

func M256TestEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256TestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is non-zero. 

FOR j := 0 to 3
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm256_test_epi64_mask'. Requires AVX512F.

func M256TestnEpi32Mask 

func M256TestnEpi32Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256TestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 7
	i := j*32
	k[j] := ((a[i+31:i] AND b[i+31:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm256_testn_epi32_mask'. Requires AVX512F.

func M256TestnEpi64Mask 

func M256TestnEpi64Mask(a x86.M256i, b x86.M256i) (dst x86.Mmask8)

M256TestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 3
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm256_testn_epi64_mask'. Requires AVX512F.

func M512AbsEpi32 

func M512AbsEpi32(a x86.M512i) (dst x86.M512i)

M512AbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ABS(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm512_abs_epi32'. Requires AVX512F.

func M512AbsEpi64 

func M512AbsEpi64(a x86.M512i) (dst x86.M512i)

M512AbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ABS(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm512_abs_epi64'. Requires AVX512F.

func M512AcosPd 

func M512AcosPd(a x86.M512d) (dst x86.M512d)

M512AcosPd: Compute the inverse cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ACOS(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_acos_pd'. Requires AVX512F.

func M512AcosPs 

func M512AcosPs(a x86.M512) (dst x86.M512)

M512AcosPs: Compute the inverse cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ACOS(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_acos_ps'. Requires AVX512F.

func M512AcoshPd 

func M512AcoshPd(a x86.M512d) (dst x86.M512d)

M512AcoshPd: Compute the inverse hyperbolic cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ACOSH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_acosh_pd'. Requires AVX512F.

func M512AcoshPs 

func M512AcoshPs(a x86.M512) (dst x86.M512)

M512AcoshPs: Compute the inverse hyperbolic cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ACOSH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_acosh_ps'. Requires AVX512F.

func M512AddEpi64 

func M512AddEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512AddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[i+63:i] + b[i+63:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm512_add_epi64'. Requires AVX512F.

func M512AlignrEpi64 

func M512AlignrEpi64(a x86.M512i, b x86.M512i, count int) (dst x86.M512i)

M512AlignrEpi64: Concatenate 'a' and 'b' into a 128-byte immediate result, shift the result right by 'count' 64-bit elements, and store the low 64 bytes (8 elements) in 'dst'. 

temp[1023:512] := a[511:0]
temp[511:0] := b[511:0]
temp[1023:0] := temp[1023:0] >> (64*count)
dst[511:0] := temp[511:0]
dst[MAX:512] := 0

Instruction: 'VALIGNQ'. Intrinsic: '_mm512_alignr_epi64'. Requires AVX512F.

func M512AsinPd 

func M512AsinPd(a x86.M512d) (dst x86.M512d)

M512AsinPd: Compute the inverse sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ASIN(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_asin_pd'. Requires AVX512F.

func M512AsinPs 

func M512AsinPs(a x86.M512) (dst x86.M512)

M512AsinPs: Compute the inverse sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ASIN(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_asin_ps'. Requires AVX512F.

func M512AsinhPd 

func M512AsinhPd(a x86.M512d) (dst x86.M512d)

M512AsinhPd: Compute the inverse hyperbolic sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ASINH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_asinh_pd'. Requires AVX512F.

func M512AsinhPs 

func M512AsinhPs(a x86.M512) (dst x86.M512)

M512AsinhPs: Compute the inverse hyperbolic sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ASINH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_asinh_ps'. Requires AVX512F.

func M512Atan2Pd 

func M512Atan2Pd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512Atan2Pd: Compute the inverse tangent of packed double-precision (64-bit) floating-point elements in 'a' divided by packed elements in 'b', and store the results in 'dst' expressed in radians. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ATAN(a[i+63:i] / b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atan2_pd'. Requires AVX512F.

func M512Atan2Ps 

func M512Atan2Ps(a x86.M512, b x86.M512) (dst x86.M512)

M512Atan2Ps: Compute the inverse tangent of packed single-precision (32-bit) floating-point elements in 'a' divided by packed elements in 'b', and store the results in 'dst' expressed in radians. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ATAN(a[i+31:i] / b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atan2_ps'. Requires AVX512F.

func M512AtanPd 

func M512AtanPd(a x86.M512d) (dst x86.M512d)

M512AtanPd: Compute the inverse tangent of packed double-precision (64-bit) floating-point elements in 'a' and store the results in 'dst' expressed in radians. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ATAN(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atan_pd'. Requires AVX512F.

func M512AtanPs 

func M512AtanPs(a x86.M512) (dst x86.M512)

M512AtanPs: Compute the inverse tangent of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' expressed in radians. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ATAN(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atan_ps'. Requires AVX512F.

func M512AtanhPd 

func M512AtanhPd(a x86.M512d) (dst x86.M512d)

M512AtanhPd: Compute the inverse hyperbolic tangent of packed double-precision (64-bit) floating-point elements in 'a' and store the results in 'dst' expressed in radians. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ATANH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atanh_pd'. Requires AVX512F.

func M512AtanhPs 

func M512AtanhPs(a x86.M512) (dst x86.M512)

M512AtanhPs: Compute the inverse hyperblic tangent of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' expressed in radians. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ATANH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_atanh_ps'. Requires AVX512F.

func M512BroadcastF32x4 

func M512BroadcastF32x4(a x86.M128) (dst x86.M512)

M512BroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	dst[i+31:i] := a[n+31:n]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm512_broadcast_f32x4'. Requires AVX512F.

func M512BroadcastF64x4 

func M512BroadcastF64x4(a x86.M256d) (dst x86.M512d)

M512BroadcastF64x4: Broadcast the 4 packed double-precision (64-bit) floating-point elements from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	dst[i+63:i] := a[n+63:n]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF64X4'. Intrinsic: '_mm512_broadcast_f64x4'. Requires AVX512F.

func M512BroadcastI32x4 

func M512BroadcastI32x4(a x86.M128i) (dst x86.M512i)

M512BroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	dst[i+31:i] := a[n+31:n]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm512_broadcast_i32x4'. Requires AVX512F.

func M512BroadcastI64x4 

func M512BroadcastI64x4(a x86.M256i) (dst x86.M512i)

M512BroadcastI64x4: Broadcast the 4 packed 64-bit integers from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	dst[i+63:i] := a[n+63:n]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI64X4'. Intrinsic: '_mm512_broadcast_i64x4'. Requires AVX512F.

func M512BroadcastdEpi32 

func M512BroadcastdEpi32(a x86.M128i) (dst x86.M512i)

M512BroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := a[31:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_broadcastd_epi32'. Requires AVX512F.

func M512BroadcastqEpi64 

func M512BroadcastqEpi64(a x86.M128i) (dst x86.M512i)

M512BroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[63:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_broadcastq_epi64'. Requires AVX512F.

func M512BroadcastsdPd 

func M512BroadcastsdPd(a x86.M128d) (dst x86.M512d)

M512BroadcastsdPd: Broadcast the low double-precision (64-bit) floating-point element from 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[63:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSD'. Intrinsic: '_mm512_broadcastsd_pd'. Requires AVX512F.

func M512BroadcastssPs 

func M512BroadcastssPs(a x86.M128) (dst x86.M512)

M512BroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := a[31:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm512_broadcastss_ps'. Requires AVX512F.

func M512Castpd128Pd512 

func M512Castpd128Pd512(a x86.M128d) (dst x86.M512d)

M512Castpd128Pd512: Cast vector of type __m128d to type __m512d; the upper 384 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castpd128_pd512'. Requires AVX512F.

func M512Castpd256Pd512 

func M512Castpd256Pd512(a x86.M256d) (dst x86.M512d)

M512Castpd256Pd512: Cast vector of type __m256d to type __m512d; the upper 256 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castpd256_pd512'. Requires AVX512F.

func M512Castpd512Pd128 

func M512Castpd512Pd128(a x86.M512d) (dst x86.M128d)

M512Castpd512Pd128: Cast vector of type __m512d to type __m128d. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castpd512_pd128'. Requires AVX512F.

func M512Castpd512Pd256 

func M512Castpd512Pd256(a x86.M512d) (dst x86.M256d)

M512Castpd512Pd256: Cast vector of type __m512d to type __m256d. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castpd512_pd256'. Requires AVX512F.

func M512Castps128Ps512 

func M512Castps128Ps512(a x86.M128) (dst x86.M512)

M512Castps128Ps512: Cast vector of type __m128 to type __m512; the upper 384 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castps128_ps512'. Requires AVX512F.

func M512Castps256Ps512 

func M512Castps256Ps512(a x86.M256) (dst x86.M512)

M512Castps256Ps512: Cast vector of type __m256 to type __m512; the upper 256 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castps256_ps512'. Requires AVX512F.

func M512Castps512Ps128 

func M512Castps512Ps128(a x86.M512) (dst x86.M128)

M512Castps512Ps128: Cast vector of type __m512 to type __m128. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castps512_ps128'. Requires AVX512F.

func M512Castps512Ps256 

func M512Castps512Ps256(a x86.M512) (dst x86.M256)

M512Castps512Ps256: Cast vector of type __m512 to type __m256. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castps512_ps256'. Requires AVX512F.

func M512Castsi128Si512 

func M512Castsi128Si512(a x86.M128i) (dst x86.M512i)

M512Castsi128Si512: Cast vector of type __m128i to type __m512i; the upper 384 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castsi128_si512'. Requires AVX512F.

func M512Castsi256Si512 

func M512Castsi256Si512(a x86.M256i) (dst x86.M512i)

M512Castsi256Si512: Cast vector of type __m256i to type __m512i; the upper 256 bits of the result are undefined. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castsi256_si512'. Requires AVX512F.

func M512Castsi512Si128 

func M512Castsi512Si128(a x86.M512i) (dst x86.M128i)

M512Castsi512Si128: Cast vector of type __m512i to type __m128i. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castsi512_si128'. Requires AVX512F.

func M512Castsi512Si256 

func M512Castsi512Si256(a x86.M512i) (dst x86.M256i)

M512Castsi512Si256: Cast vector of type __m512i to type __m256i. 

This intrinsic is only used for compilation and does not generate any

instructions, thus it has zero latency.

Instruction: ”. Intrinsic: '_mm512_castsi512_si256'. Requires AVX512F.

func M512CbrtPd 

func M512CbrtPd(a x86.M512d) (dst x86.M512d)

M512CbrtPd: Compute the cube root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := CubeRoot(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cbrt_pd'. Requires AVX512F.

func M512CbrtPs 

func M512CbrtPs(a x86.M512) (dst x86.M512)

M512CbrtPs: Compute the cube root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := CubeRoot(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cbrt_ps'. Requires AVX512F.

func M512CdfnormPd 

func M512CdfnormPd(a x86.M512d) (dst x86.M512d)

M512CdfnormPd: Compute the cumulative distribution function of packed double-precision (64-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := CDFNormal(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cdfnorm_pd'. Requires AVX512F.

func M512CdfnormPs 

func M512CdfnormPs(a x86.M512) (dst x86.M512)

M512CdfnormPs: Compute the cumulative distribution function of packed single-precision (32-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := CDFNormal(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cdfnorm_ps'. Requires AVX512F.

func M512CdfnorminvPd 

func M512CdfnorminvPd(a x86.M512d) (dst x86.M512d)

M512CdfnorminvPd: Compute the inverse cumulative distribution function of packed double-precision (64-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := InverseCDFNormal(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cdfnorminv_pd'. Requires AVX512F.

func M512CdfnorminvPs 

func M512CdfnorminvPs(a x86.M512) (dst x86.M512)

M512CdfnorminvPs: Compute the inverse cumulative distribution function of packed single-precision (32-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := InverseCDFNormal(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cdfnorminv_ps'. Requires AVX512F.

func M512CeilPd 

func M512CeilPd(a x86.M512d) (dst x86.M512d)

M512CeilPd: Round the packed double-precision (64-bit) floating-point elements in 'a' up to an integer value, and store the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := CEIL(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_ceil_pd'. Requires AVX512F.

func M512CeilPs 

func M512CeilPs(a x86.M512) (dst x86.M512)

M512CeilPs: Round the packed single-precision (32-bit) floating-point elements in 'a' up to an integer value, and store the results as packed single-precision floating-point elements in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := CEIL(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_ceil_ps'. Requires AVX512F.

func M512CmpEpi64Mask 

func M512CmpEpi64Mask(a x86.M512i, b x86.M512i, imm8 byte) (dst x86.Mmask8)

M512CmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512CmpEpu64Mask 

func M512CmpEpu64Mask(a x86.M512i, b x86.M512i, imm8 byte) (dst x86.Mmask8)

M512CmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k'. 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512CmpeqEpi64Mask 

func M512CmpeqEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPEQQ'. Intrinsic: '_mm512_cmpeq_epi64_mask'. Requires AVX512F.

func M512CmpeqEpu64Mask 

func M512CmpeqEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmpeq_epu64_mask'. Requires AVX512F.

func M512CmpgeEpi64Mask 

func M512CmpgeEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_cmpge_epi64_mask'. Requires AVX512F.

func M512CmpgeEpu64Mask 

func M512CmpgeEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmpge_epu64_mask'. Requires AVX512F.

func M512CmpgtEpi64Mask 

func M512CmpgtEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPGTQ'. Intrinsic: '_mm512_cmpgt_epi64_mask'. Requires AVX512F.

func M512CmpgtEpu64Mask 

func M512CmpgtEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmpgt_epu64_mask'. Requires AVX512F.

func M512CmpleEpi64Mask 

func M512CmpleEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_cmple_epi64_mask'. Requires AVX512F.

func M512CmpleEpu64Mask 

func M512CmpleEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmple_epu64_mask'. Requires AVX512F.

func M512CmpltEpi32Mask 

func M512CmpltEpi32Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask16)

M512CmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 15
	i := j*32
	k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
ENDFOR
k[MAX:16] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm512_cmplt_epi32_mask'. Requires AVX512F.

func M512CmpltEpi64Mask 

func M512CmpltEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_cmplt_epi64_mask'. Requires AVX512F.

func M512CmpltEpu64Mask 

func M512CmpltEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmplt_epu64_mask'. Requires AVX512F.

func M512CmpneqEpi64Mask 

func M512CmpneqEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_cmpneq_epi64_mask'. Requires AVX512F.

func M512CmpneqEpu64Mask 

func M512CmpneqEpu64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512CmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k'. 

FOR j := 0 to 7
	i := j*64
	k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_cmpneq_epu64_mask'. Requires AVX512F.

func M512CosPd 

func M512CosPd(a x86.M512d) (dst x86.M512d)

M512CosPd: Compute the cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := COS(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cos_pd'. Requires AVX512F.

func M512CosPs 

func M512CosPs(a x86.M512) (dst x86.M512)

M512CosPs: Compute the cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := COS(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cos_ps'. Requires AVX512F.

func M512CosdPd 

func M512CosdPd(a x86.M512d) (dst x86.M512d)

M512CosdPd: Compute the cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := COSD(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cosd_pd'. Requires AVX512F.

func M512CosdPs 

func M512CosdPs(a x86.M512) (dst x86.M512)

M512CosdPs: Compute the cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := COSD(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cosd_ps'. Requires AVX512F.

func M512CoshPd 

func M512CoshPd(a x86.M512d) (dst x86.M512d)

M512CoshPd: Compute the hyperbolic cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := COSH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cosh_pd'. Requires AVX512F.

func M512CoshPs 

func M512CoshPs(a x86.M512) (dst x86.M512)

M512CoshPs: Compute the hyperbolic cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := COSH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_cosh_ps'. Requires AVX512F.

func M512CvtRoundepi32Ps 

func M512CvtRoundepi32Ps(a x86.M512i, rounding int) (dst x86.M512)

M512CvtRoundepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_cvt_roundepi32_ps'. Requires AVX512F.

func M512CvtRoundepu32Ps 

func M512CvtRoundepu32Ps(a x86.M512i, rounding int) (dst x86.M512)

M512CvtRoundepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_cvt_roundepu32_ps'. Requires AVX512F.

func M512CvtRoundpdEpi32 

func M512CvtRoundpdEpi32(a x86.M512d, rounding int) (dst x86.M256i)

M512CvtRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 32*j
			k := 64*j
			dst[i+31:i] := Convert_FP64_To_Int32(a[k+63:k])
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_cvt_roundpd_epi32'. Requires AVX512F.

func M512CvtRoundpdEpu32 

func M512CvtRoundpdEpu32(a x86.M512d, rounding int) (dst x86.M256i)

M512CvtRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 32*j
			k := 64*j
			dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[k+63:k])
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_cvt_roundpd_epu32'. Requires AVX512F.

func M512CvtRoundpdPs 

func M512CvtRoundpdPs(a x86.M512d, rounding int) (dst x86.M256)

M512CvtRoundpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 32*j
			k := 64*j
			dst[i+31:i] := Convert_FP64_To_FP32(a[k+63:k])
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_cvt_roundpd_ps'. Requires AVX512F.

func M512CvtRoundphPs 

func M512CvtRoundphPs(a x86.M256i, sae int) (dst x86.M512)

M512CvtRoundphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		m := j*16
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_cvt_roundph_ps'. Requires AVX512F.

func M512CvtRoundpsEpi32 

func M512CvtRoundpsEpi32(a x86.M512, rounding int) (dst x86.M512i)

M512CvtRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_cvt_roundps_epi32'. Requires AVX512F.

func M512CvtRoundpsEpu32 

func M512CvtRoundpsEpu32(a x86.M512, rounding int) (dst x86.M512i)

M512CvtRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_cvt_roundps_epu32'. Requires AVX512F.

func M512CvtRoundpsPd 

func M512CvtRoundpsPd(a x86.M256, sae int) (dst x86.M512d)

M512CvtRoundpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 64*j
		k := 32*j
		dst[i+63:i] := Convert_FP32_To_FP64(a[k+31:k])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_cvt_roundps_pd'. Requires AVX512F.

func M512CvtRoundpsPh 

func M512CvtRoundpsPh(a x86.M512, rounding int) (dst x86.M256i)

M512CvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_cvt_roundps_ph'. Requires AVX512F.

func M512Cvtepi16Epi32 

func M512Cvtepi16Epi32(a x86.M256i) (dst x86.M512i)

M512Cvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 16*j
	dst[i+31:i] := SignExtend(a[k+15:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm512_cvtepi16_epi32'. Requires AVX512F.

func M512Cvtepi16Epi64 

func M512Cvtepi16Epi64(a x86.M128i) (dst x86.M512i)

M512Cvtepi16Epi64: Sign extend packed 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 16*j
	dst[i+63:i] := SignExtend(a[k+15:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm512_cvtepi16_epi64'. Requires AVX512F.

func M512Cvtepi32Epi16 

func M512Cvtepi32Epi16(a x86.M512i) (dst x86.M256i)

M512Cvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 16*j
	dst[k+15:k] := Truncate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm512_cvtepi32_epi16'. Requires AVX512F.

func M512Cvtepi32Epi64 

func M512Cvtepi32Epi64(a x86.M256i) (dst x86.M512i)

M512Cvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[i+63:i] := SignExtend(a[k+31:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm512_cvtepi32_epi64'. Requires AVX512F.

func M512Cvtepi32Epi8 

func M512Cvtepi32Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 8*j
	dst[k+7:k] := Truncate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm512_cvtepi32_epi8'. Requires AVX512F.

func M512Cvtepi32Pd 

func M512Cvtepi32Pd(a x86.M256i) (dst x86.M512d)

M512Cvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*32
	m := j*64
	dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm512_cvtepi32_pd'. Requires AVX512F.

func M512Cvtepi32Ps 

func M512Cvtepi32Ps(a x86.M512i) (dst x86.M512)

M512Cvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_cvtepi32_ps'. Requires AVX512F.

func M512Cvtepi64Epi16 

func M512Cvtepi64Epi16(a x86.M512i) (dst x86.M128i)

M512Cvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 16*j
	dst[k+15:k] := Truncate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm512_cvtepi64_epi16'. Requires AVX512F.

func M512Cvtepi64Epi32 

func M512Cvtepi64Epi32(a x86.M512i) (dst x86.M256i)

M512Cvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[k+31:k] := Truncate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm512_cvtepi64_epi32'. Requires AVX512F.

func M512Cvtepi64Epi8 

func M512Cvtepi64Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 8*j
	dst[k+7:k] := Truncate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm512_cvtepi64_epi8'. Requires AVX512F.

func M512Cvtepi8Epi32 

func M512Cvtepi8Epi32(a x86.M128i) (dst x86.M512i)

M512Cvtepi8Epi32: Sign extend packed 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 8*j
	dst[i+31:i] := SignExtend(a[k+7:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm512_cvtepi8_epi32'. Requires AVX512F.

func M512Cvtepi8Epi64 

func M512Cvtepi8Epi64(a x86.M128i) (dst x86.M512i)

M512Cvtepi8Epi64: Sign extend packed 8-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 8*j
	dst[i+63:i] := SignExtend(a[k+7:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm512_cvtepi8_epi64'. Requires AVX512F.

func M512Cvtepu16Epi32 

func M512Cvtepu16Epi32(a x86.M256i) (dst x86.M512i)

M512Cvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 16*j
	dst[i+31:i] := ZeroExtend(a[k+15:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm512_cvtepu16_epi32'. Requires AVX512F.

func M512Cvtepu16Epi64 

func M512Cvtepu16Epi64(a x86.M128i) (dst x86.M512i)

M512Cvtepu16Epi64: Zero extend packed unsigned 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 16*j
	dst[i+63:i] := ZeroExtend(a[k+15:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm512_cvtepu16_epi64'. Requires AVX512F.

func M512Cvtepu32Epi64 

func M512Cvtepu32Epi64(a x86.M256i) (dst x86.M512i)

M512Cvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[i+63:i] := ZeroExtend(a[k+31:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm512_cvtepu32_epi64'. Requires AVX512F.

func M512Cvtepu32Pd 

func M512Cvtepu32Pd(a x86.M256i) (dst x86.M512d)

M512Cvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	l := j*32
	dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm512_cvtepu32_pd'. Requires AVX512F.

func M512Cvtepu32Ps 

func M512Cvtepu32Ps(a x86.M512i) (dst x86.M512)

M512Cvtepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_cvtepu32_ps'. Requires AVX512F.

func M512Cvtepu8Epi32 

func M512Cvtepu8Epi32(a x86.M128i) (dst x86.M512i)

M512Cvtepu8Epi32: Zero extend packed unsigned 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 8*j
	dst[i+31:i] := ZeroExtend(a[k+7:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm512_cvtepu8_epi32'. Requires AVX512F.

func M512Cvtepu8Epi64 

func M512Cvtepu8Epi64(a x86.M128i) (dst x86.M512i)

M512Cvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 8 byte sof 'a' to packed 64-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 8*j
	dst[i+63:i] := ZeroExtend(a[k+7:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm512_cvtepu8_epi64'. Requires AVX512F.

func M512CvtpdEpi32 

func M512CvtpdEpi32(a x86.M512d) (dst x86.M256i)

M512CvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_Int32(a[k+63:k])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_cvtpd_epi32'. Requires AVX512F.

func M512CvtpdEpu32 

func M512CvtpdEpu32(a x86.M512d) (dst x86.M256i)

M512CvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[k+63:k])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_cvtpd_epu32'. Requires AVX512F.

func M512CvtpdPs 

func M512CvtpdPs(a x86.M512d) (dst x86.M256)

M512CvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_FP32(a[k+63:k])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_cvtpd_ps'. Requires AVX512F.

func M512CvtphPs 

func M512CvtphPs(a x86.M256i) (dst x86.M512)

M512CvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	m := j*16
	dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_cvtph_ps'. Requires AVX512F.

func M512CvtpsEpi32 

func M512CvtpsEpi32(a x86.M512) (dst x86.M512i)

M512CvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_cvtps_epi32'. Requires AVX512F.

func M512CvtpsEpu32 

func M512CvtpsEpu32(a x86.M512) (dst x86.M512i)

M512CvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_cvtps_epu32'. Requires AVX512F.

func M512CvtpsPd 

func M512CvtpsPd(a x86.M256) (dst x86.M512d)

M512CvtpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[i+63:i] := Convert_FP32_To_FP64(a[k+31:k])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_cvtps_pd'. Requires AVX512F.

func M512CvtpsPh 

func M512CvtpsPh(a x86.M512, rounding int) (dst x86.M256i)

M512CvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_cvtps_ph'. Requires AVX512F.

func M512Cvtsepi32Epi16 

func M512Cvtsepi32Epi16(a x86.M512i) (dst x86.M256i)

M512Cvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_Int32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm512_cvtsepi32_epi16'. Requires AVX512F.

func M512Cvtsepi32Epi8 

func M512Cvtsepi32Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_Int32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm512_cvtsepi32_epi8'. Requires AVX512F.

func M512Cvtsepi64Epi16 

func M512Cvtsepi64Epi16(a x86.M512i) (dst x86.M128i)

M512Cvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_Int64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm512_cvtsepi64_epi16'. Requires AVX512F.

func M512Cvtsepi64Epi32 

func M512Cvtsepi64Epi32(a x86.M512i) (dst x86.M256i)

M512Cvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_Int64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm512_cvtsepi64_epi32'. Requires AVX512F.

func M512Cvtsepi64Epi8 

func M512Cvtsepi64Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_Int64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm512_cvtsepi64_epi8'. Requires AVX512F.

func M512CvttRoundpdEpi32 

func M512CvttRoundpdEpi32(a x86.M512d, sae int) (dst x86.M256i)

M512CvttRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 32*i
		k := 64*j
		dst[i+31:i] := Convert_FP64_To_IntegerTruncate(a[k+63:k])
	ENDFOR
	dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_cvtt_roundpd_epi32'. Requires AVX512F.

func M512CvttRoundpdEpu32 

func M512CvttRoundpdEpu32(a x86.M512d, sae int) (dst x86.M256i)

M512CvttRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 32*i
		k := 64*j
		dst[i+31:i] := Convert_FP64_To_UnsignedIntegerTruncate(a[k+63:k])
	ENDFOR
	dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_cvtt_roundpd_epu32'. Requires AVX512F.

func M512CvttRoundpsEpi32 

func M512CvttRoundpsEpi32(a x86.M512, sae int) (dst x86.M512i)

M512CvttRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		dst[i+31:i] := Convert_FP32_To_IntegerTruncate(a[i+31:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_cvtt_roundps_epi32'. Requires AVX512F.

func M512CvttRoundpsEpu32 

func M512CvttRoundpsEpu32(a x86.M512, sae int) (dst x86.M512i)

M512CvttRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		dst[i+31:i] := Convert_FP32_To_UnsignedIntegerTruncate(a[i+31:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_cvtt_roundps_epu32'. Requires AVX512F.

func M512CvttpdEpi32 

func M512CvttpdEpi32(a x86.M512d) (dst x86.M256i)

M512CvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[k+63:k])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_cvttpd_epi32'. Requires AVX512F.

func M512CvttpdEpu32 

func M512CvttpdEpu32(a x86.M512d) (dst x86.M256i)

M512CvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 32*j
	k := 64*j
	dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[k+63:k])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_cvttpd_epu32'. Requires AVX512F.

func M512CvttpsEpi32 

func M512CvttpsEpi32(a x86.M512) (dst x86.M512i)

M512CvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_Int32_Truncate(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_cvttps_epi32'. Requires AVX512F.

func M512CvttpsEpu32 

func M512CvttpsEpu32(a x86.M512) (dst x86.M512i)

M512CvttpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := Convert_FP32_To_UnsignedInt32_Truncate(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_cvttps_epu32'. Requires AVX512F.

func M512Cvtusepi32Epi16 

func M512Cvtusepi32Epi16(a x86.M512i) (dst x86.M256i)

M512Cvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm512_cvtusepi32_epi16'. Requires AVX512F.

func M512Cvtusepi32Epi8 

func M512Cvtusepi32Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm512_cvtusepi32_epi8'. Requires AVX512F.

func M512Cvtusepi64Epi16 

func M512Cvtusepi64Epi16(a x86.M512i) (dst x86.M128i)

M512Cvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 16*j
	dst[k+15:k] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm512_cvtusepi64_epi16'. Requires AVX512F.

func M512Cvtusepi64Epi32 

func M512Cvtusepi64Epi32(a x86.M512i) (dst x86.M256i)

M512Cvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 32*j
	dst[k+31:k] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm512_cvtusepi64_epi32'. Requires AVX512F.

func M512Cvtusepi64Epi8 

func M512Cvtusepi64Epi8(a x86.M512i) (dst x86.M128i)

M512Cvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	k := 8*j
	dst[k+7:k] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm512_cvtusepi64_epi8'. Requires AVX512F.

func M512DivEpi16 

func M512DivEpi16(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpi16: Divide packed 16-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 31
	i := 16*j
	dst[i+15:i] := TRUNCATE(a[i+15:i] / b[i+15:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epi16'. Requires AVX512F.

func M512DivEpi32 

func M512DivEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpi32: Divide packed 32-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := TRUNCATE(a[i+31:i] / b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epi32'. Requires AVX512F.

func M512DivEpi64 

func M512DivEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpi64: Divide packed 64-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	dst[i+63:i] := TRUNCATE(a[i+63:i] / b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epi64'. Requires AVX512F.

func M512DivEpi8 

func M512DivEpi8(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpi8: Divide packed 8-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 63
	i := 8*j
	dst[i+7:i] := TRUNCATE(a[i+7:i] / b[i+7:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epi8'. Requires AVX512F.

func M512DivEpu16 

func M512DivEpu16(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpu16: Divide packed unsigned 16-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 31
	i := 16*j
	dst[i+15:i] := TRUNCATE(a[i+15:i] / b[i+15:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epu16'. Requires AVX512F.

func M512DivEpu32 

func M512DivEpu32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpu32: Divide packed unsigned 32-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := TRUNCATE(a[i+31:i] / b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epu32'. Requires AVX512F.

func M512DivEpu64 

func M512DivEpu64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpu64: Divide packed unsigned 64-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	dst[i+63:i] := TRUNCATE(a[i+63:i] / b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epu64'. Requires AVX512F.

func M512DivEpu8 

func M512DivEpu8(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512DivEpu8: Divide packed unsigned 8-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst'. 

FOR j := 0 to 63
	i := 8*j
	dst[i+7:i] := TRUNCATE(a[i+7:i] / b[i+7:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_div_epu8'. Requires AVX512F.

func M512DivPd 

func M512DivPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512DivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	dst[i+63:i] := a[i+63:i] / b[i+63:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_div_pd'. Requires AVX512F.

func M512DivPs 

func M512DivPs(a x86.M512, b x86.M512) (dst x86.M512)

M512DivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := a[i+31:i] / b[i+31:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_div_ps'. Requires AVX512F.

func M512DivRoundPd 

func M512DivRoundPd(a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512DivRoundPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', =and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 64*j
			dst[i+63:i] := a[i+63:i] / b[i+63:i]
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_div_round_pd'. Requires AVX512F.

func M512DivRoundPs 

func M512DivRoundPs(a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512DivRoundPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			dst[i+31:i] := a[i+31:i] / b[i+31:i]
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_div_round_ps'. Requires AVX512F.

func M512ErfPd 

func M512ErfPd(a x86.M512d) (dst x86.M512d)

M512ErfPd: Compute the error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ERF(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erf_pd'. Requires AVX512F.

func M512ErfPs 

func M512ErfPs(a x86.M512) (dst x86.M512)

M512ErfPs: Compute the error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ERF(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erf_ps'. Requires AVX512F.

func M512ErfcPd 

func M512ErfcPd(a x86.M512d) (dst x86.M512d)

M512ErfcPd: Compute the complementary error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := 1.0 - ERF(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfc_pd'. Requires AVX512F.

func M512ErfcPs 

func M512ErfcPs(a x86.M512) (dst x86.M512)

M512ErfcPs: Compute the complementary error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := 1.0 - ERF(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfc_ps'. Requires AVX512F.

func M512ErfcinvPd 

func M512ErfcinvPd(a x86.M512d) (dst x86.M512d)

M512ErfcinvPd: Compute the inverse complementary error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := 1.0 / (1.0 - ERF(a[i+63:i]))
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfcinv_pd'. Requires AVX512F.

func M512ErfcinvPs 

func M512ErfcinvPs(a x86.M512) (dst x86.M512)

M512ErfcinvPs: Compute the inverse complementary error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := 1.0 / (1.0 - ERF(a[i+31:i]))
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfcinv_ps'. Requires AVX512F.

func M512ErfinvPd 

func M512ErfinvPd(a x86.M512d) (dst x86.M512d)

M512ErfinvPd: Compute the inverse error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := 1.0 / ERF(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfinv_pd'. Requires AVX512F.

func M512ErfinvPs 

func M512ErfinvPs(a x86.M512) (dst x86.M512)

M512ErfinvPs: Compute the inverse error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := 1.0 / ERF(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_erfinv_ps'. Requires AVX512F.

func M512Exp10Pd 

func M512Exp10Pd(a x86.M512d) (dst x86.M512d)

M512Exp10Pd: Compute the exponential value of 10 raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := 10^(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp10_pd'. Requires AVX512F.

func M512Exp10Ps 

func M512Exp10Ps(a x86.M512) (dst x86.M512)

M512Exp10Ps: Compute the exponential value of 10 raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := 10^(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp10_ps'. Requires AVX512F.

func M512Exp2Pd 

func M512Exp2Pd(a x86.M512d) (dst x86.M512d)

M512Exp2Pd: Compute the exponential value of 2 raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := 2^(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp2_pd'. Requires AVX512F.

func M512Exp2Ps 

func M512Exp2Ps(a x86.M512) (dst x86.M512)

M512Exp2Ps: Compute the exponential value of 2 raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := 2^(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp2_ps'. Requires AVX512F.

func M512ExpPd 

func M512ExpPd(a x86.M512d) (dst x86.M512d)

M512ExpPd: Compute the exponential value of 'e' raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := e^(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp_pd'. Requires AVX512F.

func M512ExpPs 

func M512ExpPs(a x86.M512) (dst x86.M512)

M512ExpPs: Compute the exponential value of 'e' raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := e^(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_exp_ps'. Requires AVX512F.

func M512Expm1Pd 

func M512Expm1Pd(a x86.M512d) (dst x86.M512d)

M512Expm1Pd: Compute the exponential value of 'e' raised to the power of packed double-precision (64-bit) floating-point elements in 'a', subtract one from each element, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := e^(a[i+63:i]) - 1.0
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_expm1_pd'. Requires AVX512F.

func M512Expm1Ps 

func M512Expm1Ps(a x86.M512) (dst x86.M512)

M512Expm1Ps: Compute the exponential value of 'e' raised to the power of packed single-precision (32-bit) floating-point elements in 'a', subtract one from each element, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := e^(a[i+31:i]) - 1.0
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_expm1_ps'. Requires AVX512F.

func M512Extractf32x4Ps 

func M512Extractf32x4Ps(a x86.M512, imm8 byte) (dst x86.M128)

M512Extractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
2: dst[127:0] := a[383:256]
3: dst[127:0] := a[511:384]
ESAC
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm512_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Extractf64x4Pd 

func M512Extractf64x4Pd(a x86.M512d, imm8 byte) (dst x86.M256d)

M512Extractf64x4Pd: Extract 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[255:0] := a[255:0]
1: dst[255:0] := a[511:256]
ESAC
dst[MAX:256] := 0

Instruction: 'VEXTRACTF64X4'. Intrinsic: '_mm512_extractf64x4_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Extracti32x4Epi32 

func M512Extracti32x4Epi32(a x86.M512i, imm8 byte) (dst x86.M128i)

M512Extracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
2: dst[127:0] := a[383:256]
3: dst[127:0] := a[511:384]
ESAC
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm512_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Extracti64x4Epi64 

func M512Extracti64x4Epi64(a x86.M512i, imm8 byte) (dst x86.M256i)

M512Extracti64x4Epi64: Extract 256 bits (composed of 4 packed 64-bit integers) from 'a', selected with 'imm8', and store the result in 'dst'. 

CASE imm8[7:0] of
0: dst[255:0] := a[255:0]
1: dst[255:0] := a[511:256]
ESAC
dst[MAX:256] := 0

Instruction: 'VEXTRACTI64X4'. Intrinsic: '_mm512_extracti64x4_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512FixupimmPd 

func M512FixupimmPd(a x86.M512d, b x86.M512d, c x86.M512i, imm8 byte) (dst x86.M512d)

M512FixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512FixupimmPs 

func M512FixupimmPs(a x86.M512, b x86.M512, c x86.M512i, imm8 byte) (dst x86.M512)

M512FixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512FixupimmRoundPd 

func M512FixupimmRoundPd(a x86.M512d, b x86.M512d, c x86.M512i, imm8 byte, rounding int) (dst x86.M512d)

M512FixupimmRoundPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_fixupimm_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512FixupimmRoundPs 

func M512FixupimmRoundPs(a x86.M512, b x86.M512, c x86.M512i, imm8 byte, rounding int) (dst x86.M512)

M512FixupimmRoundPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_fixupimm_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512FloorPd 

func M512FloorPd(a x86.M512d) (dst x86.M512d)

M512FloorPd: Round the packed double-precision (64-bit) floating-point elements in 'a' down to an integer value, and store the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := FLOOR(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_floor_pd'. Requires AVX512F.

func M512FloorPs 

func M512FloorPs(a x86.M512) (dst x86.M512)

M512FloorPs: Round the packed single-precision (32-bit) floating-point elements in 'a' down to an integer value, and store the results as packed single-precision floating-point elements in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := FLOOR(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_floor_ps'. Requires AVX512F.

func M512FmaddsubPd 

func M512FmaddsubPd(a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512FmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF (j is even)
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_fmaddsub_pd'. Requires AVX512F.

func M512FmaddsubPs 

func M512FmaddsubPs(a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512FmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	IF (j is even)
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_fmaddsub_ps'. Requires AVX512F.

func M512FmaddsubRoundPd 

func M512FmaddsubRoundPd(a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512FmaddsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF (j is even)
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
			ELSE
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_fmaddsub_round_pd'. Requires AVX512F.

func M512FmaddsubRoundPs 

func M512FmaddsubRoundPs(a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512FmaddsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF (j is even)
				dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
			ELSE
				dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_fmaddsub_round_ps'. Requires AVX512F.

func M512FmsubaddPd 

func M512FmsubaddPd(a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512FmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF (j is even)
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_fmsubadd_pd'. Requires AVX512F.

func M512FmsubaddPs 

func M512FmsubaddPs(a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512FmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	IF (j is even)
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_fmsubadd_ps'. Requires AVX512F.

func M512FmsubaddRoundPd 

func M512FmsubaddRoundPd(a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512FmsubaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF (j is even)
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
			ELSE
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_fmsubadd_round_pd'. Requires AVX512F.

func M512FmsubaddRoundPs 

func M512FmsubaddRoundPs(a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512FmsubaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF (j is even)
				dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
			ELSE
				dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_fmsubadd_round_ps'. Requires AVX512F.

func M512HypotPd 

func M512HypotPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512HypotPd: Compute the length of the hypotenous of a right triangle, with the lengths of the other two sides of the triangle stored as packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SQRT(a[i+63:i]^2 + b[i+63:i]^2)
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_hypot_pd'. Requires AVX512F.

func M512HypotPs 

func M512HypotPs(a x86.M512, b x86.M512) (dst x86.M512)

M512HypotPs: Compute the length of the hypotenous of a right triangle, with the lengths of the other two sides of the triangle stored as packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SQRT(a[i+31:i]^2 + b[i+31:i]^2)
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_hypot_ps'. Requires AVX512F.

func M512Insertf32x4 

func M512Insertf32x4(a x86.M512, b x86.M128, imm8 byte) (dst x86.M512)

M512Insertf32x4: Copy 'a' to 'dst', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: dst[127:0] := b[127:0]
1: dst[255:128] := b[127:0]
2: dst[383:256] := b[127:0]
3: dst[511:384] := b[127:0]
ESAC
dst[MAX:512] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm512_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Insertf64x4 

func M512Insertf64x4(a x86.M512d, b x86.M256d, imm8 byte) (dst x86.M512d)

M512Insertf64x4: Copy 'a' to 'dst', then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[511:0] := a[511:0]
CASE (imm8[0]) of
0: dst[255:0] := b[255:0]
1: dst[511:256] := b[255:0]
ESAC
dst[MAX:512] := 0

Instruction: 'VINSERTF64X4'. Intrinsic: '_mm512_insertf64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Inserti32x4 

func M512Inserti32x4(a x86.M512i, b x86.M128i, imm8 byte) (dst x86.M512i)

M512Inserti32x4: Copy 'a' to 'dst', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: dst[127:0] := b[127:0]
1: dst[255:128] := b[127:0]
2: dst[383:256] := b[127:0]
3: dst[511:384] := b[127:0]
ESAC
dst[MAX:512] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm512_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Inserti64x4 

func M512Inserti64x4(a x86.M512i, b x86.M256i, imm8 byte) (dst x86.M512i)

M512Inserti64x4: Copy 'a' to 'dst', then insert 256 bits (composed of 4 packed 64-bit integers) from 'b' into 'dst' at the location specified by 'imm8'. 

dst[511:0] := a[511:0]
CASE (imm8[7:0]) OF
0: dst[255:0] := b[255:0]
1: dst[511:256] := b[255:0]
ESAC
dst[MAX:512] := 0

Instruction: 'VINSERTI64X4'. Intrinsic: '_mm512_inserti64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512InvsqrtPd 

func M512InvsqrtPd(a x86.M512d) (dst x86.M512d)

M512InvsqrtPd: Compute the inverse square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := InvSQRT(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_invsqrt_pd'. Requires AVX512F.

func M512InvsqrtPs 

func M512InvsqrtPs(a x86.M512) (dst x86.M512)

M512InvsqrtPs: Compute the inverse square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := InvSQRT(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_invsqrt_ps'. Requires AVX512F.

func M512Kand 

func M512Kand(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kand: Compute the bitwise AND of 16-bit masks 'a' and 'b', and store the result in 'k'. 

k[15:0] := a[15:0] AND b[15:0]
k[MAX:16] := 0

Instruction: 'KANDW'. Intrinsic: '_mm512_kand'. Requires AVX512F.

func M512Kandn 

func M512Kandn(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kandn: Compute the bitwise AND NOT of 16-bit masks 'a' and 'b', and store the result in 'k'. 

k[15:0] := (NOT a[15:0]) AND b[15:0]
k[MAX:16] := 0

Instruction: 'KANDNW'. Intrinsic: '_mm512_kandn'. Requires AVX512F.

func M512Kmov 

func M512Kmov(a x86.Mmask16) (dst x86.Mmask16)

M512Kmov: Copy 16-bit mask 'a' to 'k'. 

k[15:0] := a[15:0]
k[MAX:16] := 0

Instruction: 'KMOVW'. Intrinsic: '_mm512_kmov'. Requires AVX512F.

func M512Knot 

func M512Knot(a x86.Mmask16) (dst x86.Mmask16)

M512Knot: Compute the bitwise NOT of 16-bit mask 'a', and store the result in 'k'. 

k[15:0] := NOT a[15:0]
k[MAX:16] := 0

Instruction: 'KNOTW'. Intrinsic: '_mm512_knot'. Requires AVX512F.

func M512Kor 

func M512Kor(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kor: Compute the bitwise OR of 16-bit masks 'a' and 'b', and store the result in 'k'. 

k[15:0] := a[15:0] OR b[15:0]
k[MAX:16] := 0

Instruction: 'KORW'. Intrinsic: '_mm512_kor'. Requires AVX512F.

func M512Kortestc 

func M512Kortestc(k1 x86.Mmask16, k2 x86.Mmask16) int

M512Kortestc: Performs bitwise OR between 'k1' and 'k2', storing the result in 'dst'. CF flag is set if 'dst' consists of all 1's. 

dst[15:0] := k1[15:0] | k2[15:0]
IF PopCount(dst[15:0]) = 16
	SetCF()
FI

Instruction: 'KORTESTW'. Intrinsic: '_mm512_kortestc'. Requires AVX512F.

func M512Kortestz 

func M512Kortestz(k1 x86.Mmask16, k2 x86.Mmask16) int

M512Kortestz: Performs bitwise OR between 'k1' and 'k2', storing the result in 'dst'. ZF flag is set if 'dst' is 0. 

dst[15:0] := k1[15:0] | k2[15:0]
IF dst = 0
	SetZF()
FI

Instruction: 'KORTESTW'. Intrinsic: '_mm512_kortestz'. Requires AVX512F.

func M512Kunpackb 

func M512Kunpackb(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kunpackb: Unpack and interleave 8 bits from masks 'a' and 'b', and store the 16-bit result in 'k'. 

k[7:0] := b[7:0]
k[15:8] := a[7:0]
k[MAX:16] := 0

Instruction: 'KUNPCKBW'. Intrinsic: '_mm512_kunpackb'. Requires AVX512F.

func M512Kxnor 

func M512Kxnor(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kxnor: Compute the bitwise XNOR of 16-bit masks 'a' and 'b', and store the result in 'k'. 

k[15:0] := NOT (a[15:0] XOR b[15:0])
k[MAX:16] := 0

Instruction: 'KXNORW'. Intrinsic: '_mm512_kxnor'. Requires AVX512F.

func M512Kxor 

func M512Kxor(a x86.Mmask16, b x86.Mmask16) (dst x86.Mmask16)

M512Kxor: Compute the bitwise XOR of 16-bit masks 'a' and 'b', and store the result in 'k'. 

k[15:0] := a[15:0] XOR b[15:0]
k[MAX:16] := 0

Instruction: 'KXORW'. Intrinsic: '_mm512_kxor'. Requires AVX512F.

func M512Log10Pd 

func M512Log10Pd(a x86.M512d) (dst x86.M512d)

M512Log10Pd: Compute the base-10 logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := log10(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log10_pd'. Requires AVX512F.

func M512Log10Ps 

func M512Log10Ps(a x86.M512) (dst x86.M512)

M512Log10Ps: Compute the base-10 logarithm of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := log10(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log10_ps'. Requires AVX512F.

func M512Log1pPd 

func M512Log1pPd(a x86.M512d) (dst x86.M512d)

M512Log1pPd: Compute the natural logarithm of one plus packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ln(1.0 + a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log1p_pd'. Requires AVX512F.

func M512Log1pPs 

func M512Log1pPs(a x86.M512) (dst x86.M512)

M512Log1pPs: Compute the natural logarithm of one plus packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ln(1.0 + a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log1p_ps'. Requires AVX512F.

func M512Log2Pd 

func M512Log2Pd(a x86.M512d) (dst x86.M512d)

M512Log2Pd: Compute the base-2 logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := log2(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log2_pd'. Requires AVX512F.

func M512LogPd 

func M512LogPd(a x86.M512d) (dst x86.M512d)

M512LogPd: Compute the natural logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ln(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log_pd'. Requires AVX512F.

func M512LogPs 

func M512LogPs(a x86.M512) (dst x86.M512)

M512LogPs: Compute the natural logarithm of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ln(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_log_ps'. Requires AVX512F.

func M512LogbPd 

func M512LogbPd(a x86.M512d) (dst x86.M512d)

M512LogbPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ConvertExpFP64(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_logb_pd'. Requires AVX512F.

func M512LogbPs 

func M512LogbPs(a x86.M512) (dst x86.M512)

M512LogbPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision floating-point number representing the integer exponent, and store the results in 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := ConvertExpFP32(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_logb_ps'. Requires AVX512F.

func M512Mask2Permutex2varEpi32 

func M512Mask2Permutex2varEpi32(a x86.M512i, idx x86.M512i, k x86.Mmask16, b x86.M512i) (dst x86.M512i)

M512Mask2Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2D'. Intrinsic: '_mm512_mask2_permutex2var_epi32'. Requires AVX512F.

func M512Mask2Permutex2varEpi64 

func M512Mask2Permutex2varEpi64(a x86.M512i, idx x86.M512i, k x86.Mmask8, b x86.M512i) (dst x86.M512i)

M512Mask2Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2Q'. Intrinsic: '_mm512_mask2_permutex2var_epi64'. Requires AVX512F.

func M512Mask2Permutex2varPd 

func M512Mask2Permutex2varPd(a x86.M512d, idx x86.M512i, k x86.Mmask8, b x86.M512d) (dst x86.M512d)

M512Mask2Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set) 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PD'. Intrinsic: '_mm512_mask2_permutex2var_pd'. Requires AVX512F.

func M512Mask2Permutex2varPs 

func M512Mask2Permutex2varPs(a x86.M512, idx x86.M512i, k x86.Mmask16, b x86.M512) (dst x86.M512)

M512Mask2Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PS'. Intrinsic: '_mm512_mask2_permutex2var_ps'. Requires AVX512F.

func M512Mask3FmaddsubPd 

func M512Mask3FmaddsubPd(a x86.M512d, b x86.M512d, c x86.M512d, k x86.Mmask8) (dst x86.M512d)

M512Mask3FmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_mask3_fmaddsub_pd'. Requires AVX512F.

func M512Mask3FmaddsubPs 

func M512Mask3FmaddsubPs(a x86.M512, b x86.M512, c x86.M512, k x86.Mmask16) (dst x86.M512)

M512Mask3FmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_mask3_fmaddsub_ps'. Requires AVX512F.

func M512Mask3FmaddsubRoundPd 

func M512Mask3FmaddsubRoundPd(a x86.M512d, b x86.M512d, c x86.M512d, k x86.Mmask8, rounding int) (dst x86.M512d)

M512Mask3FmaddsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := c[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_mask3_fmaddsub_round_pd'. Requires AVX512F.

func M512Mask3FmaddsubRoundPs 

func M512Mask3FmaddsubRoundPs(a x86.M512, b x86.M512, c x86.M512, k x86.Mmask16, rounding int) (dst x86.M512)

M512Mask3FmaddsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := c[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_mask3_fmaddsub_round_ps'. Requires AVX512F.

func M512Mask3FmsubaddPd 

func M512Mask3FmsubaddPd(a x86.M512d, b x86.M512d, c x86.M512d, k x86.Mmask8) (dst x86.M512d)

M512Mask3FmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_mask3_fmsubadd_pd'. Requires AVX512F.

func M512Mask3FmsubaddPs 

func M512Mask3FmsubaddPs(a x86.M512, b x86.M512, c x86.M512, k x86.Mmask16) (dst x86.M512)

M512Mask3FmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_mask3_fmsubadd_ps'. Requires AVX512F.

func M512Mask3FmsubaddRoundPd 

func M512Mask3FmsubaddRoundPd(a x86.M512d, b x86.M512d, c x86.M512d, k x86.Mmask8, rounding int) (dst x86.M512d)

M512Mask3FmsubaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := c[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_mask3_fmsubadd_round_pd'. Requires AVX512F.

func M512Mask3FmsubaddRoundPs 

func M512Mask3FmsubaddRoundPs(a x86.M512, b x86.M512, c x86.M512, k x86.Mmask16, rounding int) (dst x86.M512)

M512Mask3FmsubaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := c[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_mask3_fmsubadd_round_ps'. Requires AVX512F.

func M512MaskAbsEpi32 

func M512MaskAbsEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm512_mask_abs_epi32'. Requires AVX512F.

func M512MaskAbsEpi64 

func M512MaskAbsEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm512_mask_abs_epi64'. Requires AVX512F.

func M512MaskAcosPd 

func M512MaskAcosPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAcosPd: Compute the inverse cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ACOS(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_acos_pd'. Requires AVX512F.

func M512MaskAcosPs 

func M512MaskAcosPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAcosPs: Compute the inverse cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ACOS(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_acos_ps'. Requires AVX512F.

func M512MaskAcoshPd 

func M512MaskAcoshPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAcoshPd: Compute the inverse hyperbolic cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ACOSH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_acosh_pd'. Requires AVX512F.

func M512MaskAcoshPs 

func M512MaskAcoshPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAcoshPs: Compute the inverse hyperbolic cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ACOSH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_acosh_ps'. Requires AVX512F.

func M512MaskAddEpi64 

func M512MaskAddEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm512_mask_add_epi64'. Requires AVX512F.

func M512MaskAlignrEpi64 

func M512MaskAlignrEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i, count int) (dst x86.M512i)

M512MaskAlignrEpi64: Concatenate 'a' and 'b' into a 128-byte immediate result, shift the result right by 'count' 64-bit elements, and store the low 64 bytes (8 elements) in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

temp[1023:512] := a[511:0]
temp[511:0] := b[511:0]
temp[1023:0] := temp[1023:0] >> (64*count)
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := temp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VALIGNQ'. Intrinsic: '_mm512_mask_alignr_epi64'. Requires AVX512F.

func M512MaskAsinPd 

func M512MaskAsinPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAsinPd: Compute the inverse sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ASIN(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_asin_pd'. Requires AVX512F.

func M512MaskAsinPs 

func M512MaskAsinPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAsinPs: Compute the inverse sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ASIN(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_asin_ps'. Requires AVX512F.

func M512MaskAsinhPd 

func M512MaskAsinhPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAsinhPd: Compute the inverse hyperbolic sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ASINH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_asinh_pd'. Requires AVX512F.

func M512MaskAsinhPs 

func M512MaskAsinhPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAsinhPs: Compute the inverse hyperbolic sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ASINH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_asinh_ps'. Requires AVX512F.

func M512MaskAtan2Pd 

func M512MaskAtan2Pd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskAtan2Pd: Compute the inverse tangent of packed double-precision (64-bit) floating-point elements in 'a' divided by packed elements in 'b', and store the results in 'dst' expressed in radians using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to
	i := j*64
	IF k[j]
		dst[i+63:i] := ATAN(a[i+63:i] / b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atan2_pd'. Requires AVX512F.

func M512MaskAtan2Ps 

func M512MaskAtan2Ps(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskAtan2Ps: Compute the inverse tangent of packed single-precision (32-bit) floating-point elements in 'a' divided by packed elements in 'b', and store the results in 'dst' expressed in radians using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ATAN(a[i+31:i] / b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atan2_ps'. Requires AVX512F.

func M512MaskAtanPd 

func M512MaskAtanPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAtanPd: Compute the inverse tangent of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' expressed in radians using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ATAN(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atan_pd'. Requires AVX512F.

func M512MaskAtanPs 

func M512MaskAtanPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAtanPs: Compute the inverse tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ATAN(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atan_ps'. Requires AVX512F.

func M512MaskAtanhPd 

func M512MaskAtanhPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskAtanhPd: Compute the inverse hyperbolic tangent of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' expressed in radians using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ATANH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atanh_pd'. Requires AVX512F.

func M512MaskAtanhPs 

func M512MaskAtanhPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskAtanhPs: Compute the inverse hyperbolic tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ATANH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_atanh_ps'. Requires AVX512F.

func M512MaskBroadcastF32x4 

func M512MaskBroadcastF32x4(src x86.M512, k x86.Mmask16, a x86.M128) (dst x86.M512)

M512MaskBroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm512_mask_broadcast_f32x4'. Requires AVX512F.

func M512MaskBroadcastF64x4 

func M512MaskBroadcastF64x4(src x86.M512d, k x86.Mmask8, a x86.M256d) (dst x86.M512d)

M512MaskBroadcastF64x4: Broadcast the 4 packed double-precision (64-bit) floating-point elements from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	IF k[j]
		dst[i+63:i] := a[n+63:n]
	ELSE
		dst[i+63:i] := src[n+63:n]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF64X4'. Intrinsic: '_mm512_mask_broadcast_f64x4'. Requires AVX512F.

func M512MaskBroadcastI32x4 

func M512MaskBroadcastI32x4(src x86.M512i, k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskBroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := src[n+31:n]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm512_mask_broadcast_i32x4'. Requires AVX512F.

func M512MaskBroadcastI64x4 

func M512MaskBroadcastI64x4(src x86.M512i, k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskBroadcastI64x4: Broadcast the 4 packed 64-bit integers from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	IF k[j]
		dst[i+63:i] := a[n+63:n]
	ELSE
		dst[i+63:i] := src[n+63:n]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI64X4'. Intrinsic: '_mm512_mask_broadcast_i64x4'. Requires AVX512F.

func M512MaskBroadcastdEpi32 

func M512MaskBroadcastdEpi32(src x86.M512i, k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_mask_broadcastd_epi32'. Requires AVX512F.

func M512MaskBroadcastqEpi64 

func M512MaskBroadcastqEpi64(src x86.M512i, k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_mask_broadcastq_epi64'. Requires AVX512F.

func M512MaskBroadcastsdPd 

func M512MaskBroadcastsdPd(src x86.M512d, k x86.Mmask8, a x86.M128d) (dst x86.M512d)

M512MaskBroadcastsdPd: Broadcast the low double-precision (64-bit) floating-point element from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSD'. Intrinsic: '_mm512_mask_broadcastsd_pd'. Requires AVX512F.

func M512MaskBroadcastssPs 

func M512MaskBroadcastssPs(src x86.M512, k x86.Mmask16, a x86.M128) (dst x86.M512)

M512MaskBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm512_mask_broadcastss_ps'. Requires AVX512F.

func M512MaskCbrtPd 

func M512MaskCbrtPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCbrtPd: Compute the cube root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := CubeRoot(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cbrt_pd'. Requires AVX512F.

func M512MaskCbrtPs 

func M512MaskCbrtPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCbrtPs: Compute the cube root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := CubeRoot(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cbrt_ps'. Requires AVX512F.

func M512MaskCdfnormPd 

func M512MaskCdfnormPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCdfnormPd: Compute the cumulative distribution function of packed double-precision (64-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := CDFNormal(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cdfnorm_pd'. Requires AVX512F.

func M512MaskCdfnormPs 

func M512MaskCdfnormPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCdfnormPs: Compute the cumulative distribution function of packed single-precision (32-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := CDFNormal(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cdfnorm_ps'. Requires AVX512F.

func M512MaskCdfnorminvPd 

func M512MaskCdfnorminvPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCdfnorminvPd: Compute the inverse cumulative distribution function of packed double-precision (64-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := InverseCDFNormal(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cdfnorminv_pd'. Requires AVX512F.

func M512MaskCdfnorminvPs 

func M512MaskCdfnorminvPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCdfnorminvPs: Compute the inverse cumulative distribution function of packed single-precision (32-bit) floating-point elements in 'a' using the normal distribution, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := InverseCDFNormal(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cdfnorminv_ps'. Requires AVX512F.

func M512MaskCeilPd 

func M512MaskCeilPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCeilPd: Round the packed double-precision (64-bit) floating-point elements in 'a' up to an integer value, and store the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := CEIL(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_ceil_pd'. Requires AVX512F.

func M512MaskCeilPs 

func M512MaskCeilPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCeilPs: Round the packed single-precision (32-bit) floating-point elements in 'a' up to an integer value, and store the results as packed single-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := CEIL(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_ceil_ps'. Requires AVX512F.

func M512MaskCmpEpi64Mask 

func M512MaskCmpEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.Mmask8)

M512MaskCmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_mask_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskCmpEpu64Mask 

func M512MaskCmpEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.Mmask8)

M512MaskCmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskCmpeqEpi64Mask 

func M512MaskCmpeqEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPEQQ'. Intrinsic: '_mm512_mask_cmpeq_epi64_mask'. Requires AVX512F.

func M512MaskCmpeqEpu64Mask 

func M512MaskCmpeqEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmpeq_epu64_mask'. Requires AVX512F.

func M512MaskCmpgeEpi64Mask 

func M512MaskCmpgeEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_mask_cmpge_epi64_mask'. Requires AVX512F.

func M512MaskCmpgeEpu64Mask 

func M512MaskCmpgeEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmpge_epu64_mask'. Requires AVX512F.

func M512MaskCmpgtEpi64Mask 

func M512MaskCmpgtEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPGTQ'. Intrinsic: '_mm512_mask_cmpgt_epi64_mask'. Requires AVX512F.

func M512MaskCmpgtEpu64Mask 

func M512MaskCmpgtEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmpgt_epu64_mask'. Requires AVX512F.

func M512MaskCmpleEpi64Mask 

func M512MaskCmpleEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_mask_cmple_epi64_mask'. Requires AVX512F.

func M512MaskCmpleEpu64Mask 

func M512MaskCmpleEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmple_epu64_mask'. Requires AVX512F.

func M512MaskCmpltEpi32Mask 

func M512MaskCmpltEpi32Mask(k1 x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.Mmask16)

M512MaskCmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:16] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm512_mask_cmplt_epi32_mask'. Requires AVX512F.

func M512MaskCmpltEpi64Mask 

func M512MaskCmpltEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_mask_cmplt_epi64_mask'. Requires AVX512F.

func M512MaskCmpltEpu64Mask 

func M512MaskCmpltEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmplt_epu64_mask'. Requires AVX512F.

func M512MaskCmpneqEpi64Mask 

func M512MaskCmpneqEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm512_mask_cmpneq_epi64_mask'. Requires AVX512F.

func M512MaskCmpneqEpu64Mask 

func M512MaskCmpneqEpu64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskCmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm512_mask_cmpneq_epu64_mask'. Requires AVX512F.

func M512MaskCompressEpi32 

func M512MaskCompressEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := src[511:m]
dst[MAX:512] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm512_mask_compress_epi32'. Requires AVX512F.

func M512MaskCompressEpi64 

func M512MaskCompressEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := src[511:m]
dst[MAX:512] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm512_mask_compress_epi64'. Requires AVX512F.

func M512MaskCompressPd 

func M512MaskCompressPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := src[511:m]
dst[MAX:512] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm512_mask_compress_pd'. Requires AVX512F.

func M512MaskCompressPs 

func M512MaskCompressPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := src[511:m]
dst[MAX:512] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm512_mask_compress_ps'. Requires AVX512F.

func M512MaskCosPd 

func M512MaskCosPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCosPd: Compute the cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := COS(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cos_pd'. Requires AVX512F.

func M512MaskCosPs 

func M512MaskCosPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCosPs: Compute the cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := COS(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cos_ps'. Requires AVX512F.

func M512MaskCosdPd 

func M512MaskCosdPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCosdPd: Compute the cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := COSD(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cosd_pd'. Requires AVX512F.

func M512MaskCosdPs 

func M512MaskCosdPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCosdPs: Compute the cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := COSD(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cosd_ps'. Requires AVX512F.

func M512MaskCoshPd 

func M512MaskCoshPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskCoshPd: Compute the hyperbolic cosine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := COSH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cosh_pd'. Requires AVX512F.

func M512MaskCoshPs 

func M512MaskCoshPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskCoshPs: Compute the hyperbolic cosine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := COSH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_cosh_ps'. Requires AVX512F.

func M512MaskCvtRoundepi32Ps 

func M512MaskCvtRoundepi32Ps(src x86.M512, k x86.Mmask16, a x86.M512i, rounding int) (dst x86.M512)

M512MaskCvtRoundepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_mask_cvt_roundepi32_ps'. Requires AVX512F.

func M512MaskCvtRoundepu32Ps 

func M512MaskCvtRoundepu32Ps(src x86.M512, k x86.Mmask16, a x86.M512i, rounding int) (dst x86.M512)

M512MaskCvtRoundepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_mask_cvt_roundepu32_ps'. Requires AVX512F.

func M512MaskCvtRoundpdEpi32 

func M512MaskCvtRoundpdEpi32(src x86.M256i, k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256i)

M512MaskCvtRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*32
			l := j*64
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_mask_cvt_roundpd_epi32'. Requires AVX512F.

func M512MaskCvtRoundpdEpu32 

func M512MaskCvtRoundpdEpu32(src x86.M256i, k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256i)

M512MaskCvtRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*32
			l := j*64
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_mask_cvt_roundpd_epu32'. Requires AVX512F.

func M512MaskCvtRoundpdPs 

func M512MaskCvtRoundpdPs(src x86.M256, k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256)

M512MaskCvtRoundpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*32
			l := j*64
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_mask_cvt_roundpd_ps'. Requires AVX512F.

func M512MaskCvtRoundphPs 

func M512MaskCvtRoundphPs(src x86.M512, k x86.Mmask16, a x86.M256i, sae int) (dst x86.M512)

M512MaskCvtRoundphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		m := j*16
		IF k[j]
			dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_mask_cvt_roundph_ps'. Requires AVX512F.

func M512MaskCvtRoundpsEpi32 

func M512MaskCvtRoundpsEpi32(src x86.M512i, k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512i)

M512MaskCvtRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_mask_cvt_roundps_epi32'. Requires AVX512F.

func M512MaskCvtRoundpsEpu32 

func M512MaskCvtRoundpsEpu32(src x86.M512i, k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512i)

M512MaskCvtRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_mask_cvt_roundps_epu32'. Requires AVX512F.

func M512MaskCvtRoundpsPd 

func M512MaskCvtRoundpsPd(src x86.M512d, k x86.Mmask8, a x86.M256, sae int) (dst x86.M512d)

M512MaskCvtRoundpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 64*j
		l := 32*j
		IF k[j]
			dst[i+63:i] := Convert_FP32_To_FP64(a[l+31:l])
		ELSE
			dst[i+63:i] := src[i+63:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_mask_cvt_roundps_pd'. Requires AVX512F.

func M512MaskCvtRoundpsPh 

func M512MaskCvtRoundpsPh(src x86.M256i, k x86.Mmask16, a x86.M512, rounding int) (dst x86.M256i)

M512MaskCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_mask_cvt_roundps_ph'. Requires AVX512F.

func M512MaskCvtepi16Epi32 

func M512MaskCvtepi16Epi32(src x86.M512i, k x86.Mmask16, a x86.M256i) (dst x86.M512i)

M512MaskCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	l := j*16
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm512_mask_cvtepi16_epi32'. Requires AVX512F.

func M512MaskCvtepi16Epi64 

func M512MaskCvtepi16Epi64(src x86.M512i, k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskCvtepi16Epi64: Sign extend packed 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm512_mask_cvtepi16_epi64'. Requires AVX512F.

func M512MaskCvtepi32Epi16 

func M512MaskCvtepi32Epi16(src x86.M256i, k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm512_mask_cvtepi32_epi16'. Requires AVX512F.

func M512MaskCvtepi32Epi64 

func M512MaskCvtepi32Epi64(src x86.M512i, k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm512_mask_cvtepi32_epi64'. Requires AVX512F.

func M512MaskCvtepi32Epi8 

func M512MaskCvtepi32Epi8(src x86.M128i, k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm512_mask_cvtepi32_epi8'. Requires AVX512F.

func M512MaskCvtepi32Pd 

func M512MaskCvtepi32Pd(src x86.M512d, k x86.Mmask8, a x86.M256i) (dst x86.M512d)

M512MaskCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := src[m+63:m]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm512_mask_cvtepi32_pd'. Requires AVX512F.

func M512MaskCvtepi32Ps 

func M512MaskCvtepi32Ps(src x86.M512, k x86.Mmask16, a x86.M512i) (dst x86.M512)

M512MaskCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_mask_cvtepi32_ps'. Requires AVX512F.

func M512MaskCvtepi64Epi16 

func M512MaskCvtepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm512_mask_cvtepi64_epi16'. Requires AVX512F.

func M512MaskCvtepi64Epi32 

func M512MaskCvtepi64Epi32(src x86.M256i, k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm512_mask_cvtepi64_epi32'. Requires AVX512F.

func M512MaskCvtepi64Epi8 

func M512MaskCvtepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm512_mask_cvtepi64_epi8'. Requires AVX512F.

func M512MaskCvtepi8Epi32 

func M512MaskCvtepi8Epi32(src x86.M512i, k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskCvtepi8Epi32: Sign extend packed 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm512_mask_cvtepi8_epi32'. Requires AVX512F.

func M512MaskCvtepi8Epi64 

func M512MaskCvtepi8Epi64(src x86.M512i, k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskCvtepi8Epi64: Sign extend packed 8-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm512_mask_cvtepi8_epi64'. Requires AVX512F.

func M512MaskCvtepu16Epi32 

func M512MaskCvtepu16Epi32(src x86.M512i, k x86.Mmask16, a x86.M256i) (dst x86.M512i)

M512MaskCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm512_mask_cvtepu16_epi32'. Requires AVX512F.

func M512MaskCvtepu16Epi64 

func M512MaskCvtepu16Epi64(src x86.M512i, k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm512_mask_cvtepu16_epi64'. Requires AVX512F.

func M512MaskCvtepu32Epi64 

func M512MaskCvtepu32Epi64(src x86.M512i, k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm512_mask_cvtepu32_epi64'. Requires AVX512F.

func M512MaskCvtepu32Pd 

func M512MaskCvtepu32Pd(src x86.M512d, k x86.Mmask8, a x86.M256i) (dst x86.M512d)

M512MaskCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm512_mask_cvtepu32_pd'. Requires AVX512F.

func M512MaskCvtepu32Ps 

func M512MaskCvtepu32Ps(src x86.M512, k x86.Mmask16, a x86.M512i) (dst x86.M512)

M512MaskCvtepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_mask_cvtepu32_ps'. Requires AVX512F.

func M512MaskCvtepu8Epi32 

func M512MaskCvtepu8Epi32(src x86.M512i, k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm512_mask_cvtepu8_epi32'. Requires AVX512F.

func M512MaskCvtepu8Epi64 

func M512MaskCvtepu8Epi64(src x86.M512i, k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm512_mask_cvtepu8_epi64'. Requires AVX512F.

func M512MaskCvtpdEpi32 

func M512MaskCvtpdEpi32(src x86.M256i, k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_mask_cvtpd_epi32'. Requires AVX512F.

func M512MaskCvtpdEpu32 

func M512MaskCvtpdEpu32(src x86.M256i, k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_mask_cvtpd_epu32'. Requires AVX512F.

func M512MaskCvtpdPs 

func M512MaskCvtpdPs(src x86.M256, k x86.Mmask8, a x86.M512d) (dst x86.M256)

M512MaskCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_mask_cvtpd_ps'. Requires AVX512F.

func M512MaskCvtphPs 

func M512MaskCvtphPs(src x86.M512, k x86.Mmask16, a x86.M256i) (dst x86.M512)

M512MaskCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_mask_cvtph_ps'. Requires AVX512F.

func M512MaskCvtpsEpi32 

func M512MaskCvtpsEpi32(src x86.M512i, k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_mask_cvtps_epi32'. Requires AVX512F.

func M512MaskCvtpsEpu32 

func M512MaskCvtpsEpu32(src x86.M512i, k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_mask_cvtps_epu32'. Requires AVX512F.

func M512MaskCvtpsPd 

func M512MaskCvtpsPd(src x86.M512d, k x86.Mmask8, a x86.M256) (dst x86.M512d)

M512MaskCvtpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := Convert_FP32_To_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_mask_cvtps_pd'. Requires AVX512F.

func M512MaskCvtpsPh 

func M512MaskCvtpsPh(src x86.M256i, k x86.Mmask16, a x86.M512, rounding int) (dst x86.M256i)

M512MaskCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_mask_cvtps_ph'. Requires AVX512F.

func M512MaskCvtsepi32Epi16 

func M512MaskCvtsepi32Epi16(src x86.M256i, k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm512_mask_cvtsepi32_epi16'. Requires AVX512F.

func M512MaskCvtsepi32Epi8 

func M512MaskCvtsepi32Epi8(src x86.M128i, k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm512_mask_cvtsepi32_epi8'. Requires AVX512F.

func M512MaskCvtsepi64Epi16 

func M512MaskCvtsepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm512_mask_cvtsepi64_epi16'. Requires AVX512F.

func M512MaskCvtsepi64Epi32 

func M512MaskCvtsepi64Epi32(src x86.M256i, k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm512_mask_cvtsepi64_epi32'. Requires AVX512F.

func M512MaskCvtsepi64Epi8 

func M512MaskCvtsepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm512_mask_cvtsepi64_epi8'. Requires AVX512F.

func M512MaskCvttRoundpdEpi32 

func M512MaskCvttRoundpdEpi32(src x86.M256i, k x86.Mmask8, a x86.M512d, sae int) (dst x86.M256i)

M512MaskCvttRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions. 

FOR j := 0 to 7
	i := 32*i
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_IntegerTruncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_mask_cvtt_roundpd_epi32'. Requires AVX512F.

func M512MaskCvttRoundpdEpu32 

func M512MaskCvttRoundpdEpu32(src x86.M256i, k x86.Mmask8, a x86.M512d, sae int) (dst x86.M256i)

M512MaskCvttRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 32*i
		l := 64*j
		IF k[j]
			dst[i+31:i] := Convert_FP64_To_UnsignedIntegerTruncate(a[l+63:l])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_mask_cvtt_roundpd_epu32'. Requires AVX512F.

func M512MaskCvttRoundpsEpi32 

func M512MaskCvttRoundpsEpi32(src x86.M512i, k x86.Mmask16, a x86.M512, sae int) (dst x86.M512i)

M512MaskCvttRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		IF k[j]
			dst[i+31:i] := Convert_FP32_To_IntegerTruncate(a[i+31:i])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_mask_cvtt_roundps_epi32'. Requires AVX512F.

func M512MaskCvttRoundpsEpu32 

func M512MaskCvttRoundpsEpu32(src x86.M512i, k x86.Mmask16, a x86.M512, sae int) (dst x86.M512i)

M512MaskCvttRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		IF k[j]
			dst[i+31:i] := Convert_FP32_To_UnsignedIntegerTruncate(a[i+31:i])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_mask_cvtt_roundps_epu32'. Requires AVX512F.

func M512MaskCvttpdEpi32 

func M512MaskCvttpdEpi32(src x86.M256i, k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_mask_cvttpd_epi32'. Requires AVX512F.

func M512MaskCvttpdEpu32 

func M512MaskCvttpdEpu32(src x86.M256i, k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_mask_cvttpd_epu32'. Requires AVX512F.

func M512MaskCvttpsEpi32 

func M512MaskCvttpsEpi32(src x86.M512i, k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_mask_cvttps_epi32'. Requires AVX512F.

func M512MaskCvttpsEpu32 

func M512MaskCvttpsEpu32(src x86.M512i, k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_mask_cvttps_epu32'. Requires AVX512F.

func M512MaskCvtusepi32Epi16 

func M512MaskCvtusepi32Epi16(src x86.M256i, k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm512_mask_cvtusepi32_epi16'. Requires AVX512F.

func M512MaskCvtusepi32Epi8 

func M512MaskCvtusepi32Epi8(src x86.M128i, k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm512_mask_cvtusepi32_epi8'. Requires AVX512F.

func M512MaskCvtusepi64Epi16 

func M512MaskCvtusepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm512_mask_cvtusepi64_epi16'. Requires AVX512F.

func M512MaskCvtusepi64Epi32 

func M512MaskCvtusepi64Epi32(src x86.M256i, k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm512_mask_cvtusepi64_epi32'. Requires AVX512F.

func M512MaskCvtusepi64Epi8 

func M512MaskCvtusepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm512_mask_cvtusepi64_epi8'. Requires AVX512F.

func M512MaskDivEpi32 

func M512MaskDivEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskDivEpi32: Divide packed 32-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := TRUNCATE(a[i+31:i] / b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_div_epi32'. Requires AVX512F.

func M512MaskDivEpu32 

func M512MaskDivEpu32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskDivEpu32: Divide packed unsigned 32-bit integers in 'a' by packed elements in 'b', and store the truncated results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := TRUNCATE(a[i+31:i] / b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_div_epu32'. Requires AVX512F.

func M512MaskDivPd 

func M512MaskDivPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_mask_div_pd'. Requires AVX512F.

func M512MaskDivPs 

func M512MaskDivPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_mask_div_ps'. Requires AVX512F.

func M512MaskDivRoundPd 

func M512MaskDivRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskDivRoundPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 64*j
			IF k[j]
				dst[i+63:i] := a[i+63:i] / b[i+63:i]
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_mask_div_round_pd'. Requires AVX512F.

func M512MaskDivRoundPs 

func M512MaskDivRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskDivRoundPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := a[i+31:i] / b[i+31:i]
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_mask_div_round_ps'. Requires AVX512F.

func M512MaskErfPd 

func M512MaskErfPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskErfPd: Compute the error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ERF(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erf_pd'. Requires AVX512F.

func M512MaskErfPs 

func M512MaskErfPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskErfPs: Compute the error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ERF(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erf_ps'. Requires AVX512F.

func M512MaskErfcPd 

func M512MaskErfcPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskErfcPd: Compute the complementary error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := 1.0 - ERF(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfc_pd'. Requires AVX512F.

func M512MaskErfcPs 

func M512MaskErfcPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskErfcPs: Compute the complementary error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := 1.0 - ERF(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfc_ps'. Requires AVX512F.

func M512MaskErfcinvPd 

func M512MaskErfcinvPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskErfcinvPd: Compute the inverse complementary error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := 1.0 / (1.0 - ERF(a[i+63:i]))
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfcinv_pd'. Requires AVX512F.

func M512MaskErfcinvPs 

func M512MaskErfcinvPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskErfcinvPs: Compute the inverse complementary error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := 1.0 / (1.0 - ERF(a[i+31:i]))
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfcinv_ps'. Requires AVX512F.

func M512MaskErfinvPd 

func M512MaskErfinvPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskErfinvPd: Compute the inverse error function of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := 1.0 / ERF(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfinv_pd'. Requires AVX512F.

func M512MaskErfinvPs 

func M512MaskErfinvPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskErfinvPs: Compute the inverse error function of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := 1.0 / ERF(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_erfinv_ps'. Requires AVX512F.

func M512MaskExp10Pd 

func M512MaskExp10Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskExp10Pd: Compute the exponential value of 10 raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := 10^(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp10_pd'. Requires AVX512F.

func M512MaskExp10Ps 

func M512MaskExp10Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskExp10Ps: Compute the exponential value of 10 raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := 10^(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp10_ps'. Requires AVX512F.

func M512MaskExp2Pd 

func M512MaskExp2Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskExp2Pd: Compute the exponential value of 2 raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := 2^(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp2_pd'. Requires AVX512F.

func M512MaskExp2Ps 

func M512MaskExp2Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskExp2Ps: Compute the exponential value of 2 raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := 2^(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp2_ps'. Requires AVX512F.

func M512MaskExpPd 

func M512MaskExpPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskExpPd: Compute the exponential value of 'e' raised to the power of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := e^(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp_pd'. Requires AVX512F.

func M512MaskExpPs 

func M512MaskExpPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskExpPs: Compute the exponential value of 'e' raised to the power of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := e^(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_exp_ps'. Requires AVX512F.

func M512MaskExpandEpi32 

func M512MaskExpandEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm512_mask_expand_epi32'. Requires AVX512F.

func M512MaskExpandEpi64 

func M512MaskExpandEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm512_mask_expand_epi64'. Requires AVX512F.

func M512MaskExpandPd 

func M512MaskExpandPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm512_mask_expand_pd'. Requires AVX512F.

func M512MaskExpandPs 

func M512MaskExpandPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm512_mask_expand_ps'. Requires AVX512F.

func M512MaskExpm1Pd 

func M512MaskExpm1Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskExpm1Pd: Compute the exponential value of 'e' raised to the power of packed double-precision (64-bit) floating-point elements in 'a', subtract one from each element, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := e^(a[i+63:i]) - 1.0
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_expm1_pd'. Requires AVX512F.

func M512MaskExpm1Ps 

func M512MaskExpm1Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskExpm1Ps: Compute the exponential value of 'e' raised to the power of packed single-precision (32-bit) floating-point elements in 'a', subtract one from each element, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := e^(a[i+31:i]) - 1.0
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_expm1_ps'. Requires AVX512F.

func M512MaskExtractf32x4Ps 

func M512MaskExtractf32x4Ps(src x86.M128, k x86.Mmask8, a x86.M512, imm8 byte) (dst x86.M128)

M512MaskExtractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[127:0] := a[127:0]
1: tmp[127:0] := a[255:128]
2: tmp[127:0] := a[383:256]
3: tmp[127:0] := a[511:384]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm512_mask_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskExtractf64x4Pd 

func M512MaskExtractf64x4Pd(src x86.M256d, k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M256d)

M512MaskExtractf64x4Pd: Extract 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[255:0] := a[255:0]
1: tmp[255:0] := a[511:256]
ESAC
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXTRACTF64X4'. Intrinsic: '_mm512_mask_extractf64x4_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskExtracti32x4Epi32 

func M512MaskExtracti32x4Epi32(src x86.M128i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M128i)

M512MaskExtracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
2: dst[127:0] := a[383:256]
3: dst[127:0] := a[511:384]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm512_mask_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskExtracti64x4Epi64 

func M512MaskExtracti64x4Epi64(src x86.M256i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M256i)

M512MaskExtracti64x4Epi64: Extract 256 bits (composed of 4 packed 64-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[255:0] := a[255:0]
1: dst[255:0] := a[511:256]
ESAC
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXTRACTI64X4'. Intrinsic: '_mm512_mask_extracti64x4_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskFixupimmPd 

func M512MaskFixupimmPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512i, imm8 byte) (dst x86.M512d)

M512MaskFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_mask_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskFixupimmPs 

func M512MaskFixupimmPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512i, imm8 byte) (dst x86.M512)

M512MaskFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_mask_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskFixupimmRoundPd 

func M512MaskFixupimmRoundPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512i, imm8 byte, rounding int) (dst x86.M512d)

M512MaskFixupimmRoundPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
			ELSE
				dst[i+63:i] := a[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_mask_fixupimm_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskFixupimmRoundPs 

func M512MaskFixupimmRoundPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512i, imm8 byte, rounding int) (dst x86.M512)

M512MaskFixupimmRoundPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
			ELSE
				dst[i+31:i] := a[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_mask_fixupimm_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskFloorPd 

func M512MaskFloorPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskFloorPd: Round the packed double-precision (64-bit) floating-point elements in 'a' down to an integer value, and store the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := FLOOR(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_floor_pd'. Requires AVX512F.

func M512MaskFloorPs 

func M512MaskFloorPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskFloorPs: Round the packed single-precision (32-bit) floating-point elements in 'a' down to an integer value, and store the results as packed single-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := FLOOR(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_floor_ps'. Requires AVX512F.

func M512MaskFmaddsubPd 

func M512MaskFmaddsubPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_mask_fmaddsub_pd'. Requires AVX512F.

func M512MaskFmaddsubPs 

func M512MaskFmaddsubPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_mask_fmaddsub_ps'. Requires AVX512F.

func M512MaskFmaddsubRoundPd 

func M512MaskFmaddsubRoundPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskFmaddsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := a[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_mask_fmaddsub_round_pd'. Requires AVX512F.

func M512MaskFmaddsubRoundPs 

func M512MaskFmaddsubRoundPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskFmaddsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := a[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_mask_fmaddsub_round_ps'. Requires AVX512F.

func M512MaskFmsubaddPd 

func M512MaskFmsubaddPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_mask_fmsubadd_pd'. Requires AVX512F.

func M512MaskFmsubaddPs 

func M512MaskFmsubaddPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_mask_fmsubadd_ps'. Requires AVX512F.

func M512MaskFmsubaddRoundPd 

func M512MaskFmsubaddRoundPd(a x86.M512d, k x86.Mmask8, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskFmsubaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := a[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_mask_fmsubadd_round_pd'. Requires AVX512F.

func M512MaskFmsubaddRoundPs 

func M512MaskFmsubaddRoundPs(a x86.M512, k x86.Mmask16, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskFmsubaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := a[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_mask_fmsubadd_round_ps'. Requires AVX512F.

func M512MaskHypotPd 

func M512MaskHypotPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskHypotPd: Compute the length of the hypotenous of a right triangle, with the lengths of the other two sides of the triangle stored as packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i]^2 + b[i+63:i]^2)
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_hypot_pd'. Requires AVX512F.

func M512MaskHypotPs 

func M512MaskHypotPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskHypotPs: Compute the length of the hypotenous of a right triangle, with the lengths of the other two sides of the triangle stored as packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i]^2 + b[i+31:i]^2)
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_hypot_ps'. Requires AVX512F.

func M512MaskInsertf32x4 

func M512MaskInsertf32x4(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M128, imm8 byte) (dst x86.M512)

M512MaskInsertf32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
2: tmp[383:256] := b[127:0]
3: tmp[511:384] := b[127:0]
ESAC
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm512_mask_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskInsertf64x4 

func M512MaskInsertf64x4(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M256d, imm8 byte) (dst x86.M512d)

M512MaskInsertf64x4: Copy 'a' to 'tmp', then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[0]) of
0: tmp[255:0] := b[255:0]
1: tmp[511:256] := b[255:0]
ESAC
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTF64X4'. Intrinsic: '_mm512_mask_insertf64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskInserti32x4 

func M512MaskInserti32x4(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M128i, imm8 byte) (dst x86.M512i)

M512MaskInserti32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
2: tmp[383:256] := b[127:0]
3: tmp[511:384] := b[127:0]
ESAC
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm512_mask_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskInserti64x4 

func M512MaskInserti64x4(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M256i, imm8 byte) (dst x86.M512i)

M512MaskInserti64x4: Copy 'a' to 'tmp', then insert 256 bits (composed of 4 packed 64-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[0]) of
0: tmp[255:0] := b[255:0]
1: tmp[511:256] := b[255:0]
ESAC
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTI64X4'. Intrinsic: '_mm512_mask_inserti64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskInvsqrtPd 

func M512MaskInvsqrtPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskInvsqrtPd: Compute the inverse square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := InvSQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_invsqrt_pd'. Requires AVX512F.

func M512MaskInvsqrtPs 

func M512MaskInvsqrtPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskInvsqrtPs: Compute the inverse square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := InvSQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_invsqrt_ps'. Requires AVX512F.

func M512MaskLog10Pd 

func M512MaskLog10Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskLog10Pd: Compute the base-10 logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := log10(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log10_pd'. Requires AVX512F.

func M512MaskLog10Ps 

func M512MaskLog10Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskLog10Ps: Compute the base-10 logarithm of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := log10(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log10_ps'. Requires AVX512F.

func M512MaskLog1pPd 

func M512MaskLog1pPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskLog1pPd: Compute the natural logarithm of one plus packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ln(1.0 + a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log1p_pd'. Requires AVX512F.

func M512MaskLog1pPs 

func M512MaskLog1pPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskLog1pPs: Compute the natural logarithm of one plus packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ln(1.0 + a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log1p_ps'. Requires AVX512F.

func M512MaskLog2Pd 

func M512MaskLog2Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskLog2Pd: Compute the base-2 logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := log2(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log2_pd'. Requires AVX512F.

func M512MaskLogPd 

func M512MaskLogPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskLogPd: Compute the natural logarithm of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ln(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log_pd'. Requires AVX512F.

func M512MaskLogPs 

func M512MaskLogPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskLogPs: Compute the natural logarithm of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ln(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_log_ps'. Requires AVX512F.

func M512MaskLogbPd 

func M512MaskLogbPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskLogbPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_logb_pd'. Requires AVX512F.

func M512MaskLogbPs 

func M512MaskLogbPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskLogbPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_logb_ps'. Requires AVX512F.

func M512MaskMaxEpi64 

func M512MaskMaxEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm512_mask_max_epi64'. Requires AVX512F.

func M512MaskMaxEpu64 

func M512MaskMaxEpu64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm512_mask_max_epu64'. Requires AVX512F.

func M512MaskMaxPd 

func M512MaskMaxPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_mask_max_pd'. Requires AVX512F.

func M512MaskMaxPs 

func M512MaskMaxPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_mask_max_ps'. Requires AVX512F.

func M512MaskMaxRoundPd 

func M512MaskMaxRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MaskMaxRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		IF k[j]
			dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
		ELSE
			dst[i+63:i] := src[i+63:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_mask_max_round_pd'. Requires AVX512F.

func M512MaskMaxRoundPs 

func M512MaskMaxRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MaskMaxRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		IF k[j]
			dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_mask_max_round_ps'. Requires AVX512F.

func M512MaskMinEpi64 

func M512MaskMinEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm512_mask_min_epi64'. Requires AVX512F.

func M512MaskMinEpu64 

func M512MaskMinEpu64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm512_mask_min_epu64'. Requires AVX512F.

func M512MaskMinPd 

func M512MaskMinPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_mask_min_pd'. Requires AVX512F.

func M512MaskMinPs 

func M512MaskMinPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_mask_min_ps'. Requires AVX512F.

func M512MaskMinRoundPd 

func M512MaskMinRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MaskMinRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		IF k[j]
			dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
		ELSE
			dst[i+63:i] := src[i+63:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_mask_min_round_pd'. Requires AVX512F.

func M512MaskMinRoundPs 

func M512MaskMinRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MaskMinRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		IF k[j]
			dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
		ELSE
			dst[i+31:i] := src[i+31:i]
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_mask_min_round_ps'. Requires AVX512F.

func M512MaskMovedupPd 

func M512MaskMovedupPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
tmp[191:128] := a[191:128]
tmp[255:192] := a[191:128]
tmp[319:256] := a[319:256]
tmp[383:320] := a[319:256]
tmp[447:384] := a[447:384]
tmp[511:448] := a[447:384]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm512_mask_movedup_pd'. Requires AVX512F.

func M512MaskMovehdupPs 

func M512MaskMovehdupPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
tmp[159:128] := a[191:160]
tmp[191:160] := a[191:160]
tmp[223:192] := a[255:224]
tmp[255:224] := a[255:224]
tmp[287:256] := a[319:288]
tmp[319:288] := a[319:288]
tmp[351:320] := a[383:352]
tmp[383:352] := a[383:352]
tmp[415:384] := a[447:416]
tmp[447:416] := a[447:416]
tmp[479:448] := a[511:480]
tmp[511:480] := a[511:480]
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm512_mask_movehdup_ps'. Requires AVX512F.

func M512MaskMoveldupPs 

func M512MaskMoveldupPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
tmp[159:128] := a[159:128]
tmp[191:160] := a[159:128]
tmp[223:192] := a[223:192]
tmp[255:224] := a[223:192]
tmp[287:256] := a[287:256]
tmp[319:288] := a[287:256]
tmp[351:320] := a[351:320]
tmp[383:352] := a[351:320]
tmp[415:384] := a[415:384]
tmp[447:416] := a[415:384]
tmp[479:448] := a[479:448]
tmp[511:480] := a[479:448]
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm512_mask_moveldup_ps'. Requires AVX512F.

func M512MaskMulEpi32 

func M512MaskMulEpi32(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm512_mask_mul_epi32'. Requires AVX512F.

func M512MaskMulEpu32 

func M512MaskMulEpu32(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm512_mask_mul_epu32'. Requires AVX512F.

func M512MaskMulloxEpi64 

func M512MaskMulloxEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskMulloxEpi64: Multiplies elements in packed 64-bit integer vectors 'a' and 'b' together, storing the lower 64 bits of the result in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_mullox_epi64'. Requires AVX512F.

func M512MaskNearbyintPd 

func M512MaskNearbyintPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskNearbyintPd: Rounds each packed double-precision (64-bit) floating-point element in 'a' to the nearest integer value and stores the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := NearbyInt(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_nearbyint_pd'. Requires AVX512F.

func M512MaskNearbyintPs 

func M512MaskNearbyintPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskNearbyintPs: Rounds each packed single-precision (32-bit) floating-point element in 'a' to the nearest integer value and stores the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := NearbyInt(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_nearbyint_ps'. Requires AVX512F.

func M512MaskPermutePd 

func M512MaskPermutePd(src x86.M512d, k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
IF (imm8[2] == 0) tmp_dst[191:128] := a[191:128]
IF (imm8[2] == 1) tmp_dst[191:128] := a[255:192]
IF (imm8[3] == 0) tmp_dst[255:192] := a[191:128]
IF (imm8[3] == 1) tmp_dst[255:192] := a[255:192]
IF (imm8[4] == 0) tmp_dst[319:256] := a[319:256]
IF (imm8[4] == 1) tmp_dst[319:256] := a[383:320]
IF (imm8[5] == 0) tmp_dst[383:320] := a[319:256]
IF (imm8[5] == 1) tmp_dst[383:320] := a[383:320]
IF (imm8[6] == 0) tmp_dst[447:384] := a[447:384]
IF (imm8[6] == 1) tmp_dst[447:384] := a[511:448]
IF (imm8[7] == 0) tmp_dst[511:448] := a[447:384]
IF (imm8[7] == 1) tmp_dst[511:448] := a[511:448]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_mask_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskPermutePs 

func M512MaskPermutePs(src x86.M512, k x86.Mmask16, a x86.M512, imm8 byte) (dst x86.M512)

M512MaskPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
tmp_dst[287:256] := SELECT4(a[383:256], imm8[1:0])
tmp_dst[319:288] := SELECT4(a[383:256], imm8[3:2])
tmp_dst[351:320] := SELECT4(a[383:256], imm8[5:4])
tmp_dst[383:352] := SELECT4(a[383:256], imm8[7:6])
tmp_dst[415:384] := SELECT4(a[511:384], imm8[1:0])
tmp_dst[447:416] := SELECT4(a[511:384], imm8[3:2])
tmp_dst[479:448] := SELECT4(a[511:384], imm8[5:4])
tmp_dst[511:480] := SELECT4(a[511:384], imm8[7:6])
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_mask_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskPermutevarPd 

func M512MaskPermutevarPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512i) (dst x86.M512d)

M512MaskPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
IF (b[129] == 0) tmp_dst[191:128] := a[191:128]
IF (b[129] == 1) tmp_dst[191:128] := a[255:192]
IF (b[193] == 0) tmp_dst[255:192] := a[191:128]
IF (b[193] == 1) tmp_dst[255:192] := a[255:192]
IF (b[257] == 0) tmp_dst[319:256] := a[319:256]
IF (b[257] == 1) tmp_dst[319:256] := a[383:320]
IF (b[321] == 0) tmp_dst[383:320] := a[319:256]
IF (b[321] == 1) tmp_dst[383:320] := a[383:320]
IF (b[385] == 0) tmp_dst[447:384] := a[447:384]
IF (b[385] == 1) tmp_dst[447:384] := a[511:448]
IF (b[449] == 0) tmp_dst[511:448] := a[447:384]
IF (b[449] == 1) tmp_dst[511:448] := a[511:448]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_mask_permutevar_pd'. Requires AVX512F.

func M512MaskPermutevarPs 

func M512MaskPermutevarPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512i) (dst x86.M512)

M512MaskPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
tmp_dst[159:128] := SELECT4(a[255:128], b[129:128])
tmp_dst[191:160] := SELECT4(a[255:128], b[161:160])
tmp_dst[223:192] := SELECT4(a[255:128], b[193:192])
tmp_dst[255:224] := SELECT4(a[255:128], b[225:224])
tmp_dst[287:256] := SELECT4(a[383:256], b[257:256])
tmp_dst[319:288] := SELECT4(a[383:256], b[289:288])
tmp_dst[351:320] := SELECT4(a[383:256], b[321:320])
tmp_dst[383:352] := SELECT4(a[383:256], b[353:352])
tmp_dst[415:384] := SELECT4(a[511:384], b[385:384])
tmp_dst[447:416] := SELECT4(a[511:384], b[417:416])
tmp_dst[479:448] := SELECT4(a[511:384], b[449:448])
tmp_dst[511:480] := SELECT4(a[511:384], b[481:480])
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_mask_permutevar_ps'. Requires AVX512F.

func M512MaskPermutex2varEpi32 

func M512MaskPermutex2varEpi32(a x86.M512i, k x86.Mmask16, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMT2D'. Intrinsic: '_mm512_mask_permutex2var_epi32'. Requires AVX512F.

func M512MaskPermutex2varEpi64 

func M512MaskPermutex2varEpi64(a x86.M512i, k x86.Mmask8, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMT2Q'. Intrinsic: '_mm512_mask_permutex2var_epi64'. Requires AVX512F.

func M512MaskPermutex2varPd 

func M512MaskPermutex2varPd(a x86.M512d, k x86.Mmask8, idx x86.M512i, b x86.M512d) (dst x86.M512d)

M512MaskPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMT2PD'. Intrinsic: '_mm512_mask_permutex2var_pd'. Requires AVX512F.

func M512MaskPermutex2varPs 

func M512MaskPermutex2varPs(a x86.M512, k x86.Mmask16, idx x86.M512i, b x86.M512) (dst x86.M512)

M512MaskPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMT2PS'. Intrinsic: '_mm512_mask_permutex2var_ps'. Requires AVX512F.

func M512MaskPermutexEpi64 

func M512MaskPermutexEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskPermutexEpi64: Shuffle 64-bit integers in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
tmp_dst[319:256] := SELECT4(a[511:256], imm8[1:0])
tmp_dst[383:320] := SELECT4(a[511:256], imm8[3:2])
tmp_dst[447:384] := SELECT4(a[511:256], imm8[5:4])
tmp_dst[511:448] := SELECT4(a[511:256], imm8[7:6])
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_mask_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskPermutexPd 

func M512MaskPermutexPd(src x86.M512d, k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskPermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
tmp_dst[319:256] := SELECT4(a[511:256], imm8[1:0])
tmp_dst[383:320] := SELECT4(a[511:256], imm8[3:2])
tmp_dst[447:384] := SELECT4(a[511:256], imm8[5:4])
tmp_dst[511:448] := SELECT4(a[511:256], imm8[7:6])
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_mask_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskPermutexvarEpi32 

func M512MaskPermutexvarEpi32(src x86.M512i, k x86.Mmask16, idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512MaskPermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm512_mask_permutexvar_epi32'. Requires AVX512F.

func M512MaskPermutexvarEpi64 

func M512MaskPermutexvarEpi64(src x86.M512i, k x86.Mmask8, idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512MaskPermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_mask_permutexvar_epi64'. Requires AVX512F.

func M512MaskPermutexvarPd 

func M512MaskPermutexvarPd(src x86.M512d, k x86.Mmask8, idx x86.M512i, a x86.M512d) (dst x86.M512d)

M512MaskPermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_mask_permutexvar_pd'. Requires AVX512F.

func M512MaskPermutexvarPs 

func M512MaskPermutexvarPs(src x86.M512, k x86.Mmask16, idx x86.M512i, a x86.M512) (dst x86.M512)

M512MaskPermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm512_mask_permutexvar_ps'. Requires AVX512F.

func M512MaskPowPd 

func M512MaskPowPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskPowPd: Compute the exponential value of packed double-precision (64-bit) floating-point elements in 'a' raised by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i])^(b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_pow_pd'. Requires AVX512F.

func M512MaskPowPs 

func M512MaskPowPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskPowPs: Compute the exponential value of packed single-precision (32-bit) floating-point elements in 'a' raised by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i])^(b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_pow_ps'. Requires AVX512F.

func M512MaskRcp14Pd 

func M512MaskRcp14Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm512_mask_rcp14_pd'. Requires AVX512F.

func M512MaskRcp14Ps 

func M512MaskRcp14Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm512_mask_rcp14_ps'. Requires AVX512F.

func M512MaskRecipPd 

func M512MaskRecipPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskRecipPd: Computes the reciprocal of packed double-precision (64-bit) floating-point elements in 'a', storing the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := (1 / a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_recip_pd'. Requires AVX512F.

func M512MaskRecipPs 

func M512MaskRecipPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskRecipPs: Computes the reciprocal of packed single-precision (32-bit) floating-point elements in 'a', storing the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := (1 / a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_recip_ps'. Requires AVX512F.

func M512MaskRemEpi32 

func M512MaskRemEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRemEpi32: Divide packed 32-bit integers in 'a' by packed elements in 'b', and store the remainders as packed 32-bit integers in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := REMAINDER(a[i+31:i] / b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_rem_epi32'. Requires AVX512F.

func M512MaskRemEpu32 

func M512MaskRemEpu32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRemEpu32: Divide packed unsigned 32-bit integers in 'a' by packed elements in 'b', and store the remainders as packed unsigned 32-bit integers in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := REMAINDER(a[i+31:i] / b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_rem_epu32'. Requires AVX512F.

func M512MaskRintPd 

func M512MaskRintPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskRintPd: Rounds the packed double-precision (64-bit) floating-point elements in 'a' to the nearest even integer value and stores the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundToNearestEven(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_rint_pd'. Requires AVX512F.

func M512MaskRintPs 

func M512MaskRintPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskRintPs: Rounds the packed single-precision (32-bit) floating-point elements in 'a' to the nearest even integer value and stores the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundToNearestEven(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_rint_ps'. Requires AVX512F.

func M512MaskRolEpi32 

func M512MaskRolEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm512_mask_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRolEpi64 

func M512MaskRolEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm512_mask_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRolvEpi32 

func M512MaskRolvEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm512_mask_rolv_epi32'. Requires AVX512F.

func M512MaskRolvEpi64 

func M512MaskRolvEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm512_mask_rolv_epi64'. Requires AVX512F.

func M512MaskRorEpi32 

func M512MaskRorEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm512_mask_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRorEpi64 

func M512MaskRorEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm512_mask_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRorvEpi32 

func M512MaskRorvEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm512_mask_rorv_epi32'. Requires AVX512F.

func M512MaskRorvEpi64 

func M512MaskRorvEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm512_mask_rorv_epi64'. Requires AVX512F.

func M512MaskRoundscalePd 

func M512MaskRoundscalePd(src x86.M512d, k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_mask_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRoundscalePs 

func M512MaskRoundscalePs(src x86.M512, k x86.Mmask16, a x86.M512, imm8 byte) (dst x86.M512)

M512MaskRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_mask_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRoundscaleRoundPd 

func M512MaskRoundscaleRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, imm8 byte, rounding int) (dst x86.M512d)

M512MaskRoundscaleRoundPd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_mask_roundscale_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRoundscaleRoundPs 

func M512MaskRoundscaleRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, imm8 byte, rounding int) (dst x86.M512)

M512MaskRoundscaleRoundPs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_mask_roundscale_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskRsqrt14Pd 

func M512MaskRsqrt14Pd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm512_mask_rsqrt14_pd'. Requires AVX512F.

func M512MaskRsqrt14Ps 

func M512MaskRsqrt14Ps(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm512_mask_rsqrt14_ps'. Requires AVX512F.

func M512MaskScalefPd 

func M512MaskScalefPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_mask_scalef_pd'. Requires AVX512F.

func M512MaskScalefPs 

func M512MaskScalefPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_mask_scalef_ps'. Requires AVX512F.

func M512MaskScalefRoundPd 

func M512MaskScalefRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskScalefRoundPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_mask_scalef_round_pd'. Requires AVX512F.

func M512MaskScalefRoundPs 

func M512MaskScalefRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskScalefRoundPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_mask_scalef_round_ps'. Requires AVX512F.

func M512MaskSet1Epi32 

func M512MaskSet1Epi32(src x86.M512i, k x86.Mmask16, a int) (dst x86.M512i)

M512MaskSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_mask_set1_epi32'. Requires AVX512F.

func M512MaskSet1Epi64 

func M512MaskSet1Epi64(src x86.M512i, k x86.Mmask8, a int64) (dst x86.M512i)

M512MaskSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_mask_set1_epi64'. Requires AVX512F.

func M512MaskShuffleF32x4 

func M512MaskShuffleF32x4(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512MaskShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm512_mask_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskShuffleF64x2 

func M512MaskShuffleF64x2(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm512_mask_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskShuffleI32x4 

func M512MaskShuffleI32x4(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm512_mask_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskShuffleI64x2 

func M512MaskShuffleI64x2(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm512_mask_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskShufflePd 

func M512MaskShufflePd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskShufflePd: Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]
tmp_dst[191:128] := (imm8[2] == 0) ? a[191:128] : a[255:192]
tmp_dst[255:192] := (imm8[3] == 0) ? b[191:128] : b[255:192]
tmp_dst[319:256] := (imm8[4] == 0) ? a[319:256] : a[383:320]
tmp_dst[383:320] := (imm8[5] == 0) ? b[319:256] : b[383:320]
tmp_dst[447:384] := (imm8[6] == 0) ? a[447:384] : a[511:448]
tmp_dst[511:448] := (imm8[7] == 0) ? b[447:384] : b[511:448]

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm512_mask_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskShufflePs 

func M512MaskShufflePs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512MaskShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(b[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(b[255:128], imm8[7:6])
tmp_dst[287:256] := SELECT4(a[383:256], imm8[1:0])
tmp_dst[319:288] := SELECT4(a[383:256], imm8[3:2])
tmp_dst[351:320] := SELECT4(b[383:256], imm8[5:4])
tmp_dst[383:352] := SELECT4(b[383:256], imm8[7:6])
tmp_dst[415:384] := SELECT4(a[511:384], imm8[1:0])
tmp_dst[447:416] := SELECT4(a[511:384], imm8[3:2])
tmp_dst[479:448] := SELECT4(b[511:384], imm8[5:4])
tmp_dst[511:480] := SELECT4(b[511:384], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm512_mask_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskSinPd 

func M512MaskSinPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSinPd: Compute the sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SIN(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sin_pd'. Requires AVX512F.

func M512MaskSinPs 

func M512MaskSinPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskSinPs: Compute the sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SIN(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sin_ps'. Requires AVX512F.

func M512MaskSincosPd 

func M512MaskSincosPd(cos_res *x86.M512d, sin_src x86.M512d, cos_src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSincosPd: Computes the sine and cosine of the packed double-precision (64-bit) floating-point elements in 'a' and stores the results of the sine computation in 'dst' and the results of the cosine computation in 'cos_res'. Elements are written to their respective locations using writemask 'k' (elements are copied from 'sin_src' or 'cos_src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SIN(a[i+63:i])
		cos_res[i+63:i] := COS(a[i+63:i])
	ELSE
		dst[i+63:i] := sin_src[i+63:i]
		cos_res[i+63:i] := cos_src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0
cos_res[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sincos_pd'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func M512MaskSincosPs 

func M512MaskSincosPs(cos_res *x86.M512, sin_src x86.M512, cos_src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskSincosPs: Computes the sine and cosine of the packed single-precision (32-bit) floating-point elements in 'a' and stores the results of the sine computation in 'dst' and the results of the cosine computation in 'cos_res'. Elements are written to their respective locations using writemask 'k' (elements are copied from 'sin_src' or 'cos_src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SIN(a[i+31:i])
		cos_res[i+31:i] := COS(a[i+31:i])
	ELSE
		dst[i+31:i] := sin_src[i+31:i]
		cos_res[i+31:i] := cos_src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0
cos_res[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sincos_ps'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func M512MaskSindPd 

func M512MaskSindPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSindPd: Compute the sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SIND(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sind_pd'. Requires AVX512F.

func M512MaskSindPs 

func M512MaskSindPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskSindPs: Compute the sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SIND(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sind_ps'. Requires AVX512F.

func M512MaskSinhPd 

func M512MaskSinhPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSinhPd: Compute the hyperbolic sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SINH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sinh_pd'. Requires AVX512F.

func M512MaskSinhPs 

func M512MaskSinhPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskSinhPs: Compute the hyperbolic sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SINH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_sinh_ps'. Requires AVX512F.

func M512MaskSllEpi32 

func M512MaskSllEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm512_mask_sll_epi32'. Requires AVX512F.

func M512MaskSllEpi64 

func M512MaskSllEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_mask_sll_epi64'. Requires AVX512F.

func M512MaskSlliEpi64 

func M512MaskSlliEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_mask_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskSllvEpi64 

func M512MaskSllvEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm512_mask_sllv_epi64'. Requires AVX512F.

func M512MaskSqrtPd 

func M512MaskSqrtPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_mask_sqrt_pd'. Requires AVX512F.

func M512MaskSqrtPs 

func M512MaskSqrtPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_mask_sqrt_ps'. Requires AVX512F.

func M512MaskSqrtRoundPd 

func M512MaskSqrtRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M512d)

M512MaskSqrtRoundPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := SQRT(a[i+63:i])
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_mask_sqrt_round_pd'. Requires AVX512F.

func M512MaskSqrtRoundPs 

func M512MaskSqrtRoundPs(src x86.M512, k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512)

M512MaskSqrtRoundPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := SQRT(a[i+31:i])
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_mask_sqrt_round_ps'. Requires AVX512F.

func M512MaskSraEpi32 

func M512MaskSraEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm512_mask_sra_epi32'. Requires AVX512F.

func M512MaskSraEpi64 

func M512MaskSraEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_mask_sra_epi64'. Requires AVX512F.

func M512MaskSraiEpi64 

func M512MaskSraiEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_mask_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskSravEpi64 

func M512MaskSravEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm512_mask_srav_epi64'. Requires AVX512F.

func M512MaskSrlEpi32 

func M512MaskSrlEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm512_mask_srl_epi32'. Requires AVX512F.

func M512MaskSrlEpi64 

func M512MaskSrlEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_mask_srl_epi64'. Requires AVX512F.

func M512MaskSrliEpi64 

func M512MaskSrliEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_mask_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskSrlvEpi64 

func M512MaskSrlvEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm512_mask_srlv_epi64'. Requires AVX512F.

func M512MaskSubEpi64 

func M512MaskSubEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm512_mask_sub_epi64'. Requires AVX512F.

func M512MaskSvmlRoundPd 

func M512MaskSvmlRoundPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskSvmlRoundPd: Round the packed double-precision (64-bit) floating-point elements in 'a' to the nearest integer value, and store the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := ROUND(a[i+63:i])
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_svml_round_pd'. Requires AVX512F.

func M512MaskTanPd 

func M512MaskTanPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskTanPd: Compute the tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := TAN(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tan_pd'. Requires AVX512F.

func M512MaskTanPs 

func M512MaskTanPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskTanPs: Compute the tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := TAN(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tan_ps'. Requires AVX512F.

func M512MaskTandPd 

func M512MaskTandPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskTandPd: Compute the tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := TAND(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tand_pd'. Requires AVX512F.

func M512MaskTandPs 

func M512MaskTandPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskTandPs: Compute the tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := TAND(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tand_ps'. Requires AVX512F.

func M512MaskTanhPd 

func M512MaskTanhPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskTanhPd: Compute the hyperbolic tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := TANH(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tanh_pd'. Requires AVX512F.

func M512MaskTanhPs 

func M512MaskTanhPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskTanhPs: Compute the hyperbolic tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := TANH(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_tanh_ps'. Requires AVX512F.

func M512MaskTernarylogicEpi32 

func M512MaskTernarylogicEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 32-bit granularity (32-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm512_mask_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskTernarylogicEpi64 

func M512MaskTernarylogicEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 64-bit granularity (64-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm512_mask_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskTestEpi64Mask 

func M512MaskTestEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskTestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is non-zero. 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm512_mask_test_epi64_mask'. Requires AVX512F.

func M512MaskTestnEpi32Mask 

func M512MaskTestnEpi32Mask(k1 x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.Mmask16)

M512MaskTestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 15
	i := j*32
	IF k1[j]
		k[j] := ((a[i+31:i] AND b[i+31:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:16] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm512_mask_testn_epi32_mask'. Requires AVX512F.

func M512MaskTestnEpi64Mask 

func M512MaskTestnEpi64Mask(k1 x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512MaskTestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 7
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm512_mask_testn_epi64_mask'. Requires AVX512F.

func M512MaskTruncPd 

func M512MaskTruncPd(src x86.M512d, k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskTruncPd: Truncate the packed double-precision (64-bit) floating-point elements in 'a', and store the results as packed double-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := TRUNCATE(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_trunc_pd'. Requires AVX512F.

func M512MaskTruncPs 

func M512MaskTruncPs(src x86.M512, k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskTruncPs: Truncate the packed single-precision (32-bit) floating-point elements in 'a', and store the results as packed single-precision floating-point elements in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := TRUNCATE(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mask_trunc_ps'. Requires AVX512F.

func M512MaskUnpackhiEpi32 

func M512MaskUnpackhiEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm512_mask_unpackhi_epi32'. Requires AVX512F.

func M512MaskUnpackhiEpi64 

func M512MaskUnpackhiEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm512_mask_unpackhi_epi64'. Requires AVX512F.

func M512MaskUnpackhiPd 

func M512MaskUnpackhiPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm512_mask_unpackhi_pd'. Requires AVX512F.

func M512MaskUnpackhiPs 

func M512MaskUnpackhiPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm512_mask_unpackhi_ps'. Requires AVX512F.

func M512MaskUnpackloEpi32 

func M512MaskUnpackloEpi32(src x86.M512i, k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm512_mask_unpacklo_epi32'. Requires AVX512F.

func M512MaskUnpackloEpi64 

func M512MaskUnpackloEpi64(src x86.M512i, k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm512_mask_unpacklo_epi64'. Requires AVX512F.

func M512MaskUnpackloPd 

func M512MaskUnpackloPd(src x86.M512d, k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm512_mask_unpacklo_pd'. Requires AVX512F.

func M512MaskUnpackloPs 

func M512MaskUnpackloPs(src x86.M512, k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm512_mask_unpacklo_ps'. Requires AVX512F.

func M512MaskzAbsEpi32 

func M512MaskzAbsEpi32(k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskzAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm512_maskz_abs_epi32'. Requires AVX512F.

func M512MaskzAbsEpi64 

func M512MaskzAbsEpi64(k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskzAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm512_maskz_abs_epi64'. Requires AVX512F.

func M512MaskzAddEpi32 

func M512MaskzAddEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAddEpi32: Add packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPADDD'. Intrinsic: '_mm512_maskz_add_epi32'. Requires AVX512F.

func M512MaskzAddEpi64 

func M512MaskzAddEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm512_maskz_add_epi64'. Requires AVX512F.

func M512MaskzAddPd 

func M512MaskzAddPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzAddPd: Add packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VADDPD'. Intrinsic: '_mm512_maskz_add_pd'. Requires AVX512F.

func M512MaskzAddPs 

func M512MaskzAddPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzAddPs: Add packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VADDPS'. Intrinsic: '_mm512_maskz_add_ps'. Requires AVX512F.

func M512MaskzAddRoundPd 

func M512MaskzAddRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskzAddRoundPd: Add packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := a[i+63:i] + b[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VADDPD'. Intrinsic: '_mm512_maskz_add_round_pd'. Requires AVX512F.

func M512MaskzAddRoundPs 

func M512MaskzAddRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskzAddRoundPs: Add packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := a[i+31:i] + b[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VADDPS'. Intrinsic: '_mm512_maskz_add_round_ps'. Requires AVX512F.

func M512MaskzAlignrEpi32 

func M512MaskzAlignrEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i, count int) (dst x86.M512i)

M512MaskzAlignrEpi32: Concatenate 'a' and 'b' into a 128-byte immediate result, shift the result right by 'count' 32-bit elements, and stores the low 64 bytes (16 elements) in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

temp[1023:512] := a[511:0]
temp[511:0] := b[511:0]
temp[1023:0] := temp[1023:0] >> (32*count)
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := temp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VALIGND'. Intrinsic: '_mm512_maskz_alignr_epi32'. Requires AVX512F.

func M512MaskzAlignrEpi64 

func M512MaskzAlignrEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i, count int) (dst x86.M512i)

M512MaskzAlignrEpi64: Concatenate 'a' and 'b' into a 128-byte immediate result, shift the result right by 'count' 64-bit elements, and stores the low 64 bytes (8 elements) in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

temp[1023:512] := a[511:0]
temp[511:0] := b[511:0]
temp[1023:0] := temp[1023:0] >> (64*count)
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := temp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VALIGNQ'. Intrinsic: '_mm512_maskz_alignr_epi64'. Requires AVX512F.

func M512MaskzAndEpi32 

func M512MaskzAndEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAndEpi32: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPANDD'. Intrinsic: '_mm512_maskz_and_epi32'. Requires AVX512F.

func M512MaskzAndEpi64 

func M512MaskzAndEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAndEpi64: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPANDQ'. Intrinsic: '_mm512_maskz_and_epi64'. Requires AVX512F.

func M512MaskzAndnotEpi32 

func M512MaskzAndnotEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAndnotEpi32: Compute the bitwise AND NOT of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := (NOT a[i+31:i]) AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPANDND'. Intrinsic: '_mm512_maskz_andnot_epi32'. Requires AVX512F.

func M512MaskzAndnotEpi64 

func M512MaskzAndnotEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzAndnotEpi64: Compute the bitwise AND NOT of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := (NOT a[i+63:i]) AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPANDNQ'. Intrinsic: '_mm512_maskz_andnot_epi64'. Requires AVX512F.

func M512MaskzBroadcastF32x4 

func M512MaskzBroadcastF32x4(k x86.Mmask16, a x86.M128) (dst x86.M512)

M512MaskzBroadcastF32x4: Broadcast the 4 packed single-precision (32-bit) floating-point elements from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF32X4'. Intrinsic: '_mm512_maskz_broadcast_f32x4'. Requires AVX512F.

func M512MaskzBroadcastF64x4 

func M512MaskzBroadcastF64x4(k x86.Mmask8, a x86.M256d) (dst x86.M512d)

M512MaskzBroadcastF64x4: Broadcast the 4 packed double-precision (64-bit) floating-point elements from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	IF k[j]
		dst[i+63:i] := a[n+63:n]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTF64X4'. Intrinsic: '_mm512_maskz_broadcast_f64x4'. Requires AVX512F.

func M512MaskzBroadcastI32x4 

func M512MaskzBroadcastI32x4(k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskzBroadcastI32x4: Broadcast the 4 packed 32-bit integers from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	n := (j mod 4)*32
	IF k[j]
		dst[i+31:i] := a[n+31:n]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI32X4'. Intrinsic: '_mm512_maskz_broadcast_i32x4'. Requires AVX512F.

func M512MaskzBroadcastI64x4 

func M512MaskzBroadcastI64x4(k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskzBroadcastI64x4: Broadcast the 4 packed 64-bit integers from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	n := (j mod 4)*64
	IF k[j]
		dst[i+63:i] := a[n+63:n]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTI64X4'. Intrinsic: '_mm512_maskz_broadcast_i64x4'. Requires AVX512F.

func M512MaskzBroadcastdEpi32 

func M512MaskzBroadcastdEpi32(k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskzBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_maskz_broadcastd_epi32'. Requires AVX512F.

func M512MaskzBroadcastqEpi64 

func M512MaskzBroadcastqEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskzBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_maskz_broadcastq_epi64'. Requires AVX512F.

func M512MaskzBroadcastsdPd 

func M512MaskzBroadcastsdPd(k x86.Mmask8, a x86.M128d) (dst x86.M512d)

M512MaskzBroadcastsdPd: Broadcast the low double-precision (64-bit) floating-point element from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSD'. Intrinsic: '_mm512_maskz_broadcastsd_pd'. Requires AVX512F.

func M512MaskzBroadcastssPs 

func M512MaskzBroadcastssPs(k x86.Mmask16, a x86.M128) (dst x86.M512)

M512MaskzBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm512_maskz_broadcastss_ps'. Requires AVX512F.

func M512MaskzCompressEpi32 

func M512MaskzCompressEpi32(k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskzCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := 0
dst[MAX:512] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm512_maskz_compress_epi32'. Requires AVX512F.

func M512MaskzCompressEpi64 

func M512MaskzCompressEpi64(k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskzCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := 0
dst[MAX:512] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm512_maskz_compress_epi64'. Requires AVX512F.

func M512MaskzCompressPd 

func M512MaskzCompressPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := 0
dst[MAX:512] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm512_maskz_compress_pd'. Requires AVX512F.

func M512MaskzCompressPs 

func M512MaskzCompressPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[511:m] := 0
dst[MAX:512] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm512_maskz_compress_ps'. Requires AVX512F.

func M512MaskzCvtRoundepi32Ps 

func M512MaskzCvtRoundepi32Ps(k x86.Mmask16, a x86.M512i, rounding int) (dst x86.M512)

M512MaskzCvtRoundepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_maskz_cvt_roundepi32_ps'. Requires AVX512F.

func M512MaskzCvtRoundepu32Ps 

func M512MaskzCvtRoundepu32Ps(k x86.Mmask16, a x86.M512i, rounding int) (dst x86.M512)

M512MaskzCvtRoundepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_maskz_cvt_roundepu32_ps'. Requires AVX512F.

func M512MaskzCvtRoundpdEpi32 

func M512MaskzCvtRoundpdEpi32(k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256i)

M512MaskzCvtRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 32*j
			l := 64*j
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_maskz_cvt_roundpd_epi32'. Requires AVX512F.

func M512MaskzCvtRoundpdEpu32 

func M512MaskzCvtRoundpdEpu32(k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256i)

M512MaskzCvtRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 32*j
			l := 64*j
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_maskz_cvt_roundpd_epu32'. Requires AVX512F.

func M512MaskzCvtRoundpdPs 

func M512MaskzCvtRoundpdPs(k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M256)

M512MaskzCvtRoundpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*32
			l := j*64
			IF k[j]
				dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_maskz_cvt_roundpd_ps'. Requires AVX512F.

func M512MaskzCvtRoundphPs 

func M512MaskzCvtRoundphPs(k x86.Mmask16, a x86.M256i, sae int) (dst x86.M512)

M512MaskzCvtRoundphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		m := j*16
		IF k[j]
			dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_maskz_cvt_roundph_ps'. Requires AVX512F.

func M512MaskzCvtRoundpsEpi32 

func M512MaskzCvtRoundpsEpi32(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512i)

M512MaskzCvtRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_maskz_cvt_roundps_epi32'. Requires AVX512F.

func M512MaskzCvtRoundpsEpu32 

func M512MaskzCvtRoundpsEpu32(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512i)

M512MaskzCvtRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_maskz_cvt_roundps_epu32'. Requires AVX512F.

func M512MaskzCvtRoundpsPd 

func M512MaskzCvtRoundpsPd(k x86.Mmask8, a x86.M256, sae int) (dst x86.M512d)

M512MaskzCvtRoundpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 64*j
		l := 32*j
		IF k[j]
			dst[i+63:i] := Convert_FP32_To_FP64(a[l+31:l])
		ELSE
			dst[i+63:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_maskz_cvt_roundps_pd'. Requires AVX512F.

func M512MaskzCvtRoundpsPh 

func M512MaskzCvtRoundpsPh(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M256i)

M512MaskzCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_maskz_cvt_roundps_ph'. Requires AVX512F.

func M512MaskzCvtepi16Epi32 

func M512MaskzCvtepi16Epi32(k x86.Mmask16, a x86.M256i) (dst x86.M512i)

M512MaskzCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm512_maskz_cvtepi16_epi32'. Requires AVX512F.

func M512MaskzCvtepi16Epi64 

func M512MaskzCvtepi16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepi16Epi64: Sign extend packed 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm512_maskz_cvtepi16_epi64'. Requires AVX512F.

func M512MaskzCvtepi32Epi16 

func M512MaskzCvtepi32Epi16(k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskzCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm512_maskz_cvtepi32_epi16'. Requires AVX512F.

func M512MaskzCvtepi32Epi64 

func M512MaskzCvtepi32Epi64(k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskzCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm512_maskz_cvtepi32_epi64'. Requires AVX512F.

func M512MaskzCvtepi32Epi8 

func M512MaskzCvtepi32Epi8(k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskzCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm512_maskz_cvtepi32_epi8'. Requires AVX512F.

func M512MaskzCvtepi32Pd 

func M512MaskzCvtepi32Pd(k x86.Mmask8, a x86.M256i) (dst x86.M512d)

M512MaskzCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm512_maskz_cvtepi32_pd'. Requires AVX512F.

func M512MaskzCvtepi32Ps 

func M512MaskzCvtepi32Ps(k x86.Mmask16, a x86.M512i) (dst x86.M512)

M512MaskzCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm512_maskz_cvtepi32_ps'. Requires AVX512F.

func M512MaskzCvtepi64Epi16 

func M512MaskzCvtepi64Epi16(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm512_maskz_cvtepi64_epi16'. Requires AVX512F.

func M512MaskzCvtepi64Epi32 

func M512MaskzCvtepi64Epi32(k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskzCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm512_maskz_cvtepi64_epi32'. Requires AVX512F.

func M512MaskzCvtepi64Epi8 

func M512MaskzCvtepi64Epi8(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm512_maskz_cvtepi64_epi8'. Requires AVX512F.

func M512MaskzCvtepi8Epi32 

func M512MaskzCvtepi8Epi32(k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepi8Epi32: Sign extend packed 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm512_maskz_cvtepi8_epi32'. Requires AVX512F.

func M512MaskzCvtepi8Epi64 

func M512MaskzCvtepi8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepi8Epi64: Sign extend packed 8-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm512_maskz_cvtepi8_epi64'. Requires AVX512F.

func M512MaskzCvtepu16Epi32 

func M512MaskzCvtepu16Epi32(k x86.Mmask16, a x86.M256i) (dst x86.M512i)

M512MaskzCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm512_maskz_cvtepu16_epi32'. Requires AVX512F.

func M512MaskzCvtepu16Epi64 

func M512MaskzCvtepu16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm512_maskz_cvtepu16_epi64'. Requires AVX512F.

func M512MaskzCvtepu32Epi64 

func M512MaskzCvtepu32Epi64(k x86.Mmask8, a x86.M256i) (dst x86.M512i)

M512MaskzCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm512_maskz_cvtepu32_epi64'. Requires AVX512F.

func M512MaskzCvtepu32Pd 

func M512MaskzCvtepu32Pd(k x86.Mmask8, a x86.M256i) (dst x86.M512d)

M512MaskzCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm512_maskz_cvtepu32_pd'. Requires AVX512F.

func M512MaskzCvtepu32Ps 

func M512MaskzCvtepu32Ps(k x86.Mmask16, a x86.M512i) (dst x86.M512)

M512MaskzCvtepu32Ps: Convert packed unsigned 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := ConvertUnsignedInt32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTUDQ2PS'. Intrinsic: '_mm512_maskz_cvtepu32_ps'. Requires AVX512F.

func M512MaskzCvtepu8Epi32 

func M512MaskzCvtepu8Epi32(k x86.Mmask16, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm512_maskz_cvtepu8_epi32'. Requires AVX512F.

func M512MaskzCvtepu8Epi64 

func M512MaskzCvtepu8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M512i)

M512MaskzCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 8 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm512_maskz_cvtepu8_epi64'. Requires AVX512F.

func M512MaskzCvtpdEpi32 

func M512MaskzCvtpdEpi32(k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskzCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm512_maskz_cvtpd_epi32'. Requires AVX512F.

func M512MaskzCvtpdEpu32 

func M512MaskzCvtpdEpu32(k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskzCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm512_maskz_cvtpd_epu32'. Requires AVX512F.

func M512MaskzCvtpdPs 

func M512MaskzCvtpdPs(k x86.Mmask8, a x86.M512d) (dst x86.M256)

M512MaskzCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm512_maskz_cvtpd_ps'. Requires AVX512F.

func M512MaskzCvtphPs 

func M512MaskzCvtphPs(k x86.Mmask16, a x86.M256i) (dst x86.M512)

M512MaskzCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm512_maskz_cvtph_ps'. Requires AVX512F.

func M512MaskzCvtpsEpi32 

func M512MaskzCvtpsEpi32(k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskzCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm512_maskz_cvtps_epi32'. Requires AVX512F.

func M512MaskzCvtpsEpu32 

func M512MaskzCvtpsEpu32(k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskzCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm512_maskz_cvtps_epu32'. Requires AVX512F.

func M512MaskzCvtpsPd 

func M512MaskzCvtpsPd(k x86.Mmask8, a x86.M256) (dst x86.M512d)

M512MaskzCvtpsPd: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := Convert_FP32_To_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTPS2PD'. Intrinsic: '_mm512_maskz_cvtps_pd'. Requires AVX512F.

func M512MaskzCvtpsPh 

func M512MaskzCvtpsPh(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M256i)

M512MaskzCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:256] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm512_maskz_cvtps_ph'. Requires AVX512F.

func M512MaskzCvtsepi32Epi16 

func M512MaskzCvtsepi32Epi16(k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskzCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm512_maskz_cvtsepi32_epi16'. Requires AVX512F.

func M512MaskzCvtsepi32Epi8 

func M512MaskzCvtsepi32Epi8(k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskzCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm512_maskz_cvtsepi32_epi8'. Requires AVX512F.

func M512MaskzCvtsepi64Epi16 

func M512MaskzCvtsepi64Epi16(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm512_maskz_cvtsepi64_epi16'. Requires AVX512F.

func M512MaskzCvtsepi64Epi32 

func M512MaskzCvtsepi64Epi32(k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskzCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm512_maskz_cvtsepi64_epi32'. Requires AVX512F.

func M512MaskzCvtsepi64Epi8 

func M512MaskzCvtsepi64Epi8(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm512_maskz_cvtsepi64_epi8'. Requires AVX512F.

func M512MaskzCvttRoundpdEpi32 

func M512MaskzCvttRoundpdEpi32(k x86.Mmask8, a x86.M512d, sae int) (dst x86.M256i)

M512MaskzCvttRoundpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 32*i
		l := 64*j
		IF k[j]
			dst[i+31:i] := Convert_FP64_To_IntegerTruncate(a[l+63:l])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_maskz_cvtt_roundpd_epi32'. Requires AVX512F.

func M512MaskzCvttRoundpdEpu32 

func M512MaskzCvttRoundpdEpu32(k x86.Mmask8, a x86.M512d, sae int) (dst x86.M256i)

M512MaskzCvttRoundpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := 32*i
		l := 64*j
		IF k[j]
			dst[i+31:i] := Convert_FP64_To_UnsignedIntegerTruncate(a[l+63:l])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_maskz_cvtt_roundpd_epu32'. Requires AVX512F.

func M512MaskzCvttRoundpsEpi32 

func M512MaskzCvttRoundpsEpi32(k x86.Mmask16, a x86.M512, sae int) (dst x86.M512i)

M512MaskzCvttRoundpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		IF k[j]
			dst[i+31:i] := Convert_FP32_To_IntegerTruncate(a[i+31:i])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_maskz_cvtt_roundps_epi32'. Requires AVX512F.

func M512MaskzCvttRoundpsEpu32 

func M512MaskzCvttRoundpsEpu32(k x86.Mmask16, a x86.M512, sae int) (dst x86.M512i)

M512MaskzCvttRoundpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := 32*i
		IF k[j]
			dst[i+31:i] := Convert_FP32_To_UnsignedIntegerTruncate(a[i+31:i])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_maskz_cvtt_roundps_epu32'. Requires AVX512F.

func M512MaskzCvttpdEpi32 

func M512MaskzCvttpdEpi32(k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskzCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm512_maskz_cvttpd_epi32'. Requires AVX512F.

func M512MaskzCvttpdEpu32 

func M512MaskzCvttpdEpu32(k x86.Mmask8, a x86.M512d) (dst x86.M256i)

M512MaskzCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm512_maskz_cvttpd_epu32'. Requires AVX512F.

func M512MaskzCvttpsEpi32 

func M512MaskzCvttpsEpi32(k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskzCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm512_maskz_cvttps_epi32'. Requires AVX512F.

func M512MaskzCvttpsEpu32 

func M512MaskzCvttpsEpu32(k x86.Mmask16, a x86.M512) (dst x86.M512i)

M512MaskzCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm512_maskz_cvttps_epu32'. Requires AVX512F.

func M512MaskzCvtusepi32Epi16 

func M512MaskzCvtusepi32Epi16(k x86.Mmask16, a x86.M512i) (dst x86.M256i)

M512MaskzCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm512_maskz_cvtusepi32_epi16'. Requires AVX512F.

func M512MaskzCvtusepi32Epi8 

func M512MaskzCvtusepi32Epi8(k x86.Mmask16, a x86.M512i) (dst x86.M128i)

M512MaskzCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm512_maskz_cvtusepi32_epi8'. Requires AVX512F.

func M512MaskzCvtusepi64Epi16 

func M512MaskzCvtusepi64Epi16(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm512_maskz_cvtusepi64_epi16'. Requires AVX512F.

func M512MaskzCvtusepi64Epi32 

func M512MaskzCvtusepi64Epi32(k x86.Mmask8, a x86.M512i) (dst x86.M256i)

M512MaskzCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm512_maskz_cvtusepi64_epi32'. Requires AVX512F.

func M512MaskzCvtusepi64Epi8 

func M512MaskzCvtusepi64Epi8(k x86.Mmask8, a x86.M512i) (dst x86.M128i)

M512MaskzCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm512_maskz_cvtusepi64_epi8'. Requires AVX512F.

func M512MaskzDivPd 

func M512MaskzDivPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_maskz_div_pd'. Requires AVX512F.

func M512MaskzDivPs 

func M512MaskzDivPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_maskz_div_ps'. Requires AVX512F.

func M512MaskzDivRoundPd 

func M512MaskzDivRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskzDivRoundPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 64*j
			IF k[j]
				dst[i+63:i] := a[i+63:i] / b[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm512_maskz_div_round_pd'. Requires AVX512F.

func M512MaskzDivRoundPs 

func M512MaskzDivRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskzDivRoundPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := 32*j
			IF k[j]
				dst[i+31:i] := a[i+31:i] / b[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm512_maskz_div_round_ps'. Requires AVX512F.

func M512MaskzExpandEpi32 

func M512MaskzExpandEpi32(k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskzExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm512_maskz_expand_epi32'. Requires AVX512F.

func M512MaskzExpandEpi64 

func M512MaskzExpandEpi64(k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskzExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm512_maskz_expand_epi64'. Requires AVX512F.

func M512MaskzExpandPd 

func M512MaskzExpandPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm512_maskz_expand_pd'. Requires AVX512F.

func M512MaskzExpandPs 

func M512MaskzExpandPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm512_maskz_expand_ps'. Requires AVX512F.

func M512MaskzExtractf32x4Ps 

func M512MaskzExtractf32x4Ps(k x86.Mmask8, a x86.M512, imm8 byte) (dst x86.M128)

M512MaskzExtractf32x4Ps: Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[127:0] := a[127:0]
1: tmp[127:0] := a[255:128]
2: tmp[127:0] := a[383:256]
3: tmp[127:0] := a[511:384]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTF32X4'. Intrinsic: '_mm512_maskz_extractf32x4_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzExtractf64x4Pd 

func M512MaskzExtractf64x4Pd(k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M256d)

M512MaskzExtractf64x4Pd: Extract 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: tmp[255:0] := a[255:0]
1: tmp[255:0] := a[511:256]
ESAC
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXTRACTF64X4'. Intrinsic: '_mm512_maskz_extractf64x4_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzExtracti32x4Epi32 

func M512MaskzExtracti32x4Epi32(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M128i)

M512MaskzExtracti32x4Epi32: Extract 128 bits (composed of 4 packed 32-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[127:0] := a[127:0]
1: dst[127:0] := a[255:128]
2: dst[127:0] := a[383:256]
3: dst[127:0] := a[511:384]
ESAC
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXTRACTI32X4'. Intrinsic: '_mm512_maskz_extracti32x4_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzExtracti64x4Epi64 

func M512MaskzExtracti64x4Epi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M256i)

M512MaskzExtracti64x4Epi64: Extract 256 bits (composed of 4 packed 64-bit integers) from 'a', selected with 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE imm8[7:0] of
0: dst[255:0] := a[255:0]
1: dst[255:0] := a[511:256]
ESAC
FOR j := 0 to 3
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:256] := 0

Instruction: 'VEXTRACTI64X4'. Intrinsic: '_mm512_maskz_extracti64x4_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzFixupimmPd 

func M512MaskzFixupimmPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512i, imm8 byte) (dst x86.M512d)

M512MaskzFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_maskz_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzFixupimmPs 

func M512MaskzFixupimmPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512i, imm8 byte) (dst x86.M512)

M512MaskzFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_maskz_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzFixupimmRoundPd 

func M512MaskzFixupimmRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512i, imm8 byte, rounding int) (dst x86.M512d)

M512MaskzFixupimmRoundPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm512_maskz_fixupimm_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzFixupimmRoundPs 

func M512MaskzFixupimmRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512i, imm8 byte, rounding int) (dst x86.M512)

M512MaskzFixupimmRoundPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm512_maskz_fixupimm_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzFmaddPd 

func M512MaskzFmaddPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm512_maskz_fmadd_pd'. Requires AVX512F.

func M512MaskzFmaddPs 

func M512MaskzFmaddPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm512_maskz_fmadd_ps'. Requires AVX512F.

func M512MaskzFmaddRoundPd 

func M512MaskzFmaddRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFmaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm512_maskz_fmadd_round_pd'. Requires AVX512F.

func M512MaskzFmaddRoundPs 

func M512MaskzFmaddRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFmaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'a' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				a[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm512_maskz_fmadd_round_ps'. Requires AVX512F.

func M512MaskzFmaddsubPd 

func M512MaskzFmaddsubPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_maskz_fmaddsub_pd'. Requires AVX512F.

func M512MaskzFmaddsubPs 

func M512MaskzFmaddsubPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_maskz_fmaddsub_ps'. Requires AVX512F.

func M512MaskzFmaddsubRoundPd 

func M512MaskzFmaddsubRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFmaddsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm512_maskz_fmaddsub_round_pd'. Requires AVX512F.

func M512MaskzFmaddsubRoundPs 

func M512MaskzFmaddsubRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFmaddsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm512_maskz_fmaddsub_round_ps'. Requires AVX512F.

func M512MaskzFmsubPd 

func M512MaskzFmsubPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm512_maskz_fmsub_pd'. Requires AVX512F.

func M512MaskzFmsubPs 

func M512MaskzFmsubPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm512_maskz_fmsub_ps'. Requires AVX512F.

func M512MaskzFmsubRoundPd 

func M512MaskzFmsubRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFmsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm512_maskz_fmsub_round_pd'. Requires AVX512F.

func M512MaskzFmsubRoundPs 

func M512MaskzFmsubRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFmsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm512_maskz_fmsub_round_ps'. Requires AVX512F.

func M512MaskzFmsubaddPd 

func M512MaskzFmsubaddPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_maskz_fmsubadd_pd'. Requires AVX512F.

func M512MaskzFmsubaddPs 

func M512MaskzFmsubaddPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_maskz_fmsubadd_ps'. Requires AVX512F.

func M512MaskzFmsubaddRoundPd 

func M512MaskzFmsubaddRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFmsubaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				IF (j is even)
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
				ELSE
					dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
				FI
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm512_maskz_fmsubadd_round_pd'. Requires AVX512F.

func M512MaskzFmsubaddRoundPs 

func M512MaskzFmsubaddRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFmsubaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				IF (j is even)
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
				ELSE
					dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
				FI
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm512_maskz_fmsubadd_round_ps'. Requires AVX512F.

func M512MaskzFnmaddPd 

func M512MaskzFnmaddPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm512_maskz_fnmadd_pd'. Requires AVX512F.

func M512MaskzFnmaddPs 

func M512MaskzFnmaddPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm512_maskz_fnmadd_ps'. Requires AVX512F.

func M512MaskzFnmaddRoundPd 

func M512MaskzFnmaddRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFnmaddRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm512_maskz_fnmadd_round_pd'. Requires AVX512F.

func M512MaskzFnmaddRoundPs 

func M512MaskzFnmaddRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFnmaddRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm512_maskz_fnmadd_round_ps'. Requires AVX512F.

func M512MaskzFnmsubPd 

func M512MaskzFnmsubPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d) (dst x86.M512d)

M512MaskzFnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm512_maskz_fnmsub_pd'. Requires AVX512F.

func M512MaskzFnmsubPs 

func M512MaskzFnmsubPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512) (dst x86.M512)

M512MaskzFnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm512_maskz_fnmsub_ps'. Requires AVX512F.

func M512MaskzFnmsubRoundPd 

func M512MaskzFnmsubRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, c x86.M512d, rounding int) (dst x86.M512d)

M512MaskzFnmsubRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). Rounding is done according to the 'rounding' parameter, which can be one of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm512_maskz_fnmsub_round_pd'. Requires AVX512F.

func M512MaskzFnmsubRoundPs 

func M512MaskzFnmsubRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, c x86.M512, rounding int) (dst x86.M512)

M512MaskzFnmsubRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm512_maskz_fnmsub_round_ps'. Requires AVX512F.

func M512MaskzGetexpPd 

func M512MaskzGetexpPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzGetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm512_maskz_getexp_pd'. Requires AVX512F.

func M512MaskzGetexpPs 

func M512MaskzGetexpPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzGetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm512_maskz_getexp_ps'. Requires AVX512F.

func M512MaskzGetexpRoundPd 

func M512MaskzGetexpRoundPd(k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M512d)

M512MaskzGetexpRoundPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := ConvertExpFP64(a[i+63:i])
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm512_maskz_getexp_round_pd'. Requires AVX512F.

func M512MaskzGetexpRoundPs 

func M512MaskzGetexpRoundPs(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512)

M512MaskzGetexpRoundPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := ConvertExpFP32(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm512_maskz_getexp_round_ps'. Requires AVX512F.

func M512MaskzGetmantPd 

func M512MaskzGetmantPd(k x86.Mmask8, a x86.M512d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M512d)

M512MaskzGetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm512_maskz_getmant_pd'. Requires AVX512F.

func M512MaskzGetmantPs 

func M512MaskzGetmantPs(k x86.Mmask16, a x86.M512, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M512)

M512MaskzGetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm512_maskz_getmant_ps'. Requires AVX512F.

func M512MaskzGetmantRoundPd 

func M512MaskzGetmantRoundPd(k x86.Mmask8, a x86.M512d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M512d)

M512MaskzGetmantRoundPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm512_maskz_getmant_round_pd'. Requires AVX512F.

func M512MaskzGetmantRoundPs 

func M512MaskzGetmantRoundPs(k x86.Mmask16, a x86.M512, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M512)

M512MaskzGetmantRoundPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm512_maskz_getmant_round_ps'. Requires AVX512F.

func M512MaskzInsertf32x4 

func M512MaskzInsertf32x4(k x86.Mmask16, a x86.M512, b x86.M128, imm8 byte) (dst x86.M512)

M512MaskzInsertf32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
2: tmp[383:256] := b[127:0]
3: tmp[511:384] := b[127:0]
ESAC
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTF32X4'. Intrinsic: '_mm512_maskz_insertf32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzInsertf64x4 

func M512MaskzInsertf64x4(k x86.Mmask8, a x86.M512d, b x86.M256d, imm8 byte) (dst x86.M512d)

M512MaskzInsertf64x4: Copy 'a' to 'tmp', then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[0]) of
0: tmp[255:0] := b[255:0]
1: tmp[511:256] := b[255:0]
ESAC
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTF64X4'. Intrinsic: '_mm512_maskz_insertf64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzInserti32x4 

func M512MaskzInserti32x4(k x86.Mmask16, a x86.M512i, b x86.M128i, imm8 byte) (dst x86.M512i)

M512MaskzInserti32x4: Copy 'a' to 'tmp', then insert 128 bits (composed of 4 packed 32-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[1:0]) of
0: tmp[127:0] := b[127:0]
1: tmp[255:128] := b[127:0]
2: tmp[383:256] := b[127:0]
3: tmp[511:384] := b[127:0]
ESAC
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTI32X4'. Intrinsic: '_mm512_maskz_inserti32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzInserti64x4 

func M512MaskzInserti64x4(k x86.Mmask8, a x86.M512i, b x86.M256i, imm8 byte) (dst x86.M512i)

M512MaskzInserti64x4: Copy 'a' to 'tmp', then insert 256 bits (composed of 4 packed 64-bit integers) from 'b' into 'tmp' at the location specified by 'imm8'. Store 'tmp' to 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[511:0] := a[511:0]
CASE (imm8[0]) of
0: tmp[255:0] := b[255:0]
1: tmp[511:256] := b[255:0]
ESAC
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VINSERTI64X4'. Intrinsic: '_mm512_maskz_inserti64x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzMaxEpi32 

func M512MaskzMaxEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMaxEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXSD'. Intrinsic: '_mm512_maskz_max_epi32'. Requires AVX512F.

func M512MaskzMaxEpi64 

func M512MaskzMaxEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm512_maskz_max_epi64'. Requires AVX512F.

func M512MaskzMaxEpu32 

func M512MaskzMaxEpu32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMaxEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXUD'. Intrinsic: '_mm512_maskz_max_epu32'. Requires AVX512F.

func M512MaskzMaxEpu64 

func M512MaskzMaxEpu64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm512_maskz_max_epu64'. Requires AVX512F.

func M512MaskzMaxPd 

func M512MaskzMaxPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_maskz_max_pd'. Requires AVX512F.

func M512MaskzMaxPs 

func M512MaskzMaxPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_maskz_max_ps'. Requires AVX512F.

func M512MaskzMaxRoundPd 

func M512MaskzMaxRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MaskzMaxRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		IF k[j]
			dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
		ELSE
			dst[i+63:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_maskz_max_round_pd'. Requires AVX512F.

func M512MaskzMaxRoundPs 

func M512MaskzMaxRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MaskzMaxRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		IF k[j]
			dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_maskz_max_round_ps'. Requires AVX512F.

func M512MaskzMinEpi32 

func M512MaskzMinEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMinEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINSD'. Intrinsic: '_mm512_maskz_min_epi32'. Requires AVX512F.

func M512MaskzMinEpi64 

func M512MaskzMinEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm512_maskz_min_epi64'. Requires AVX512F.

func M512MaskzMinEpu32 

func M512MaskzMinEpu32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMinEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINUD'. Intrinsic: '_mm512_maskz_min_epu32'. Requires AVX512F.

func M512MaskzMinEpu64 

func M512MaskzMinEpu64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm512_maskz_min_epu64'. Requires AVX512F.

func M512MaskzMinPd 

func M512MaskzMinPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_maskz_min_pd'. Requires AVX512F.

func M512MaskzMinPs 

func M512MaskzMinPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_maskz_min_ps'. Requires AVX512F.

func M512MaskzMinRoundPd 

func M512MaskzMinRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MaskzMinRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		IF k[j]
			dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
		ELSE
			dst[i+63:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_maskz_min_round_pd'. Requires AVX512F.

func M512MaskzMinRoundPs 

func M512MaskzMinRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MaskzMinRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		IF k[j]
			dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
		ELSE
			dst[i+31:i] := 0
		FI
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_maskz_min_round_ps'. Requires AVX512F.

func M512MaskzMovEpi32 

func M512MaskzMovEpi32(k x86.Mmask16, a x86.M512i) (dst x86.M512i)

M512MaskzMovEpi32: Move packed 32-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVDQA32'. Intrinsic: '_mm512_maskz_mov_epi32'. Requires AVX512F.

func M512MaskzMovEpi64 

func M512MaskzMovEpi64(k x86.Mmask8, a x86.M512i) (dst x86.M512i)

M512MaskzMovEpi64: Move packed 64-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVDQA64'. Intrinsic: '_mm512_maskz_mov_epi64'. Requires AVX512F.

func M512MaskzMovPd 

func M512MaskzMovPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzMovPd: Move packed double-precision (64-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVAPD'. Intrinsic: '_mm512_maskz_mov_pd'. Requires AVX512F.

func M512MaskzMovPs 

func M512MaskzMovPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzMovPs: Move packed single-precision (32-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVAPS'. Intrinsic: '_mm512_maskz_mov_ps'. Requires AVX512F.

func M512MaskzMovedupPd 

func M512MaskzMovedupPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
tmp[191:128] := a[191:128]
tmp[255:192] := a[191:128]
tmp[319:256] := a[319:256]
tmp[383:320] := a[319:256]
tmp[447:384] := a[447:384]
tmp[511:448] := a[447:384]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm512_maskz_movedup_pd'. Requires AVX512F.

func M512MaskzMovehdupPs 

func M512MaskzMovehdupPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
tmp[159:128] := a[191:160]
tmp[191:160] := a[191:160]
tmp[223:192] := a[255:224]
tmp[255:224] := a[255:224]
tmp[287:256] := a[319:288]
tmp[319:288] := a[319:288]
tmp[351:320] := a[383:352]
tmp[383:352] := a[383:352]
tmp[415:384] := a[447:416]
tmp[447:416] := a[447:416]
tmp[479:448] := a[511:480]
tmp[511:480] := a[511:480]
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm512_maskz_movehdup_ps'. Requires AVX512F.

func M512MaskzMoveldupPs 

func M512MaskzMoveldupPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
tmp[159:128] := a[159:128]
tmp[191:160] := a[159:128]
tmp[223:192] := a[223:192]
tmp[255:224] := a[223:192]
tmp[287:256] := a[287:256]
tmp[319:288] := a[287:256]
tmp[351:320] := a[351:320]
tmp[383:352] := a[351:320]
tmp[415:384] := a[415:384]
tmp[447:416] := a[415:384]
tmp[479:448] := a[479:448]
tmp[511:480] := a[479:448]
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm512_maskz_moveldup_ps'. Requires AVX512F.

func M512MaskzMulEpi32 

func M512MaskzMulEpi32(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm512_maskz_mul_epi32'. Requires AVX512F.

func M512MaskzMulEpu32 

func M512MaskzMulEpu32(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm512_maskz_mul_epu32'. Requires AVX512F.

func M512MaskzMulPd 

func M512MaskzMulPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzMulPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm512_maskz_mul_pd'. Requires AVX512F.

func M512MaskzMulPs 

func M512MaskzMulPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzMulPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm512_maskz_mul_ps'. Requires AVX512F.

func M512MaskzMulRoundPd 

func M512MaskzMulRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskzMulRoundPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := a[i+63:i] * b[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm512_maskz_mul_round_pd'. Requires AVX512F.

func M512MaskzMulRoundPs 

func M512MaskzMulRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskzMulRoundPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := a[i+31:i] * b[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm512_maskz_mul_round_ps'. Requires AVX512F.

func M512MaskzMulloEpi32 

func M512MaskzMulloEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzMulloEpi32: Multiply the packed 32-bit integers in 'a' and 'b', producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		tmp[63:0] := a[i+31:i] * b[i+31:i]
		dst[i+31:i] := tmp[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULLD'. Intrinsic: '_mm512_maskz_mullo_epi32'. Requires AVX512F.

func M512MaskzOrEpi32 

func M512MaskzOrEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzOrEpi32: Compute the bitwise OR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] OR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPORD'. Intrinsic: '_mm512_maskz_or_epi32'. Requires AVX512F.

func M512MaskzOrEpi64 

func M512MaskzOrEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzOrEpi64: Compute the bitwise OR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] OR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPORQ'. Intrinsic: '_mm512_maskz_or_epi64'. Requires AVX512F.

func M512MaskzPermutePd 

func M512MaskzPermutePd(k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskzPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
IF (imm8[2] == 0) tmp_dst[191:128] := a[191:128]
IF (imm8[2] == 1) tmp_dst[191:128] := a[255:192]
IF (imm8[3] == 0) tmp_dst[255:192] := a[191:128]
IF (imm8[3] == 1) tmp_dst[255:192] := a[255:192]
IF (imm8[4] == 0) tmp_dst[319:256] := a[319:256]
IF (imm8[4] == 1) tmp_dst[319:256] := a[383:320]
IF (imm8[5] == 0) tmp_dst[383:320] := a[319:256]
IF (imm8[5] == 1) tmp_dst[383:320] := a[383:320]
IF (imm8[6] == 0) tmp_dst[447:384] := a[447:384]
IF (imm8[6] == 1) tmp_dst[447:384] := a[511:448]
IF (imm8[7] == 0) tmp_dst[511:448] := a[447:384]
IF (imm8[7] == 1) tmp_dst[511:448] := a[511:448]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_maskz_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzPermutePs 

func M512MaskzPermutePs(k x86.Mmask16, a x86.M512, imm8 byte) (dst x86.M512)

M512MaskzPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
tmp_dst[287:256] := SELECT4(a[383:256], imm8[1:0])
tmp_dst[319:288] := SELECT4(a[383:256], imm8[3:2])
tmp_dst[351:320] := SELECT4(a[383:256], imm8[5:4])
tmp_dst[383:352] := SELECT4(a[383:256], imm8[7:6])
tmp_dst[415:384] := SELECT4(a[511:384], imm8[1:0])
tmp_dst[447:416] := SELECT4(a[511:384], imm8[3:2])
tmp_dst[479:448] := SELECT4(a[511:384], imm8[5:4])
tmp_dst[511:480] := SELECT4(a[511:384], imm8[7:6])
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_maskz_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzPermutevarPd 

func M512MaskzPermutevarPd(k x86.Mmask8, a x86.M512d, b x86.M512i) (dst x86.M512d)

M512MaskzPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
IF (b[129] == 0) tmp_dst[191:128] := a[191:128]
IF (b[129] == 1) tmp_dst[191:128] := a[255:192]
IF (b[193] == 0) tmp_dst[255:192] := a[191:128]
IF (b[193] == 1) tmp_dst[255:192] := a[255:192]
IF (b[257] == 0) tmp_dst[319:256] := a[319:256]
IF (b[257] == 1) tmp_dst[319:256] := a[383:320]
IF (b[321] == 0) tmp_dst[383:320] := a[319:256]
IF (b[321] == 1) tmp_dst[383:320] := a[383:320]
IF (b[385] == 0) tmp_dst[447:384] := a[447:384]
IF (b[385] == 1) tmp_dst[447:384] := a[511:448]
IF (b[449] == 0) tmp_dst[511:448] := a[447:384]
IF (b[449] == 1) tmp_dst[511:448] := a[511:448]
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_maskz_permutevar_pd'. Requires AVX512F.

func M512MaskzPermutevarPs 

func M512MaskzPermutevarPs(k x86.Mmask16, a x86.M512, b x86.M512i) (dst x86.M512)

M512MaskzPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
tmp_dst[159:128] := SELECT4(a[255:128], b[129:128])
tmp_dst[191:160] := SELECT4(a[255:128], b[161:160])
tmp_dst[223:192] := SELECT4(a[255:128], b[193:192])
tmp_dst[255:224] := SELECT4(a[255:128], b[225:224])
tmp_dst[287:256] := SELECT4(a[383:256], b[257:256])
tmp_dst[319:288] := SELECT4(a[383:256], b[289:288])
tmp_dst[351:320] := SELECT4(a[383:256], b[321:320])
tmp_dst[383:352] := SELECT4(a[383:256], b[353:352])
tmp_dst[415:384] := SELECT4(a[511:384], b[385:384])
tmp_dst[447:416] := SELECT4(a[511:384], b[417:416])
tmp_dst[479:448] := SELECT4(a[511:384], b[449:448])
tmp_dst[511:480] := SELECT4(a[511:384], b[481:480])
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_maskz_permutevar_ps'. Requires AVX512F.

func M512MaskzPermutex2varEpi32 

func M512MaskzPermutex2varEpi32(k x86.Mmask16, a x86.M512i, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+4]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm512_maskz_permutex2var_epi32'. Requires AVX512F.

func M512MaskzPermutex2varEpi64 

func M512MaskzPermutex2varEpi64(k x86.Mmask8, a x86.M512i, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+3]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm512_maskz_permutex2var_epi64'. Requires AVX512F.

func M512MaskzPermutex2varPd 

func M512MaskzPermutex2varPd(k x86.Mmask8, a x86.M512d, idx x86.M512i, b x86.M512d) (dst x86.M512d)

M512MaskzPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+3]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm512_maskz_permutex2var_pd'. Requires AVX512F.

func M512MaskzPermutex2varPs 

func M512MaskzPermutex2varPs(k x86.Mmask16, a x86.M512, idx x86.M512i, b x86.M512) (dst x86.M512)

M512MaskzPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+4]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm512_maskz_permutex2var_ps'. Requires AVX512F.

func M512MaskzPermutexEpi64 

func M512MaskzPermutexEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzPermutexEpi64: Shuffle 64-bit integers in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
tmp_dst[319:256] := SELECT4(a[511:256], imm8[1:0])
tmp_dst[383:320] := SELECT4(a[511:256], imm8[3:2])
tmp_dst[447:384] := SELECT4(a[511:256], imm8[5:4])
tmp_dst[511:448] := SELECT4(a[511:256], imm8[7:6])
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_maskz_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzPermutexPd 

func M512MaskzPermutexPd(k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskzPermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

tmp_dst[63:0] := SELECT4(a[255:0], imm8[1:0])
tmp_dst[127:64] := SELECT4(a[255:0], imm8[3:2])
tmp_dst[191:128] := SELECT4(a[255:0], imm8[5:4])
tmp_dst[255:192] := SELECT4(a[255:0], imm8[7:6])
tmp_dst[319:256] := SELECT4(a[511:256], imm8[1:0])
tmp_dst[383:320] := SELECT4(a[511:256], imm8[3:2])
tmp_dst[447:384] := SELECT4(a[511:256], imm8[5:4])
tmp_dst[511:448] := SELECT4(a[511:256], imm8[7:6])
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_maskz_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzPermutexvarEpi32 

func M512MaskzPermutexvarEpi32(k x86.Mmask16, idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512MaskzPermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm512_maskz_permutexvar_epi32'. Requires AVX512F.

func M512MaskzPermutexvarEpi64 

func M512MaskzPermutexvarEpi64(k x86.Mmask8, idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512MaskzPermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_maskz_permutexvar_epi64'. Requires AVX512F.

func M512MaskzPermutexvarPd 

func M512MaskzPermutexvarPd(k x86.Mmask8, idx x86.M512i, a x86.M512d) (dst x86.M512d)

M512MaskzPermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	IF k[j]
		dst[i+63:i] := a[id+63:id]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_maskz_permutexvar_pd'. Requires AVX512F.

func M512MaskzPermutexvarPs 

func M512MaskzPermutexvarPs(k x86.Mmask16, idx x86.M512i, a x86.M512) (dst x86.M512)

M512MaskzPermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	IF k[j]
		dst[i+31:i] := a[id+31:id]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm512_maskz_permutexvar_ps'. Requires AVX512F.

func M512MaskzRcp14Pd 

func M512MaskzRcp14Pd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm512_maskz_rcp14_pd'. Requires AVX512F.

func M512MaskzRcp14Ps 

func M512MaskzRcp14Ps(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm512_maskz_rcp14_ps'. Requires AVX512F.

func M512MaskzRolEpi32 

func M512MaskzRolEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm512_maskz_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRolEpi64 

func M512MaskzRolEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm512_maskz_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRolvEpi32 

func M512MaskzRolvEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm512_maskz_rolv_epi32'. Requires AVX512F.

func M512MaskzRolvEpi64 

func M512MaskzRolvEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm512_maskz_rolv_epi64'. Requires AVX512F.

func M512MaskzRorEpi32 

func M512MaskzRorEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm512_maskz_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRorEpi64 

func M512MaskzRorEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm512_maskz_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRorvEpi32 

func M512MaskzRorvEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm512_maskz_rorv_epi32'. Requires AVX512F.

func M512MaskzRorvEpi64 

func M512MaskzRorvEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm512_maskz_rorv_epi64'. Requires AVX512F.

func M512MaskzRoundscalePd 

func M512MaskzRoundscalePd(k x86.Mmask8, a x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskzRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_maskz_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRoundscalePs 

func M512MaskzRoundscalePs(k x86.Mmask16, a x86.M512, imm8 byte) (dst x86.M512)

M512MaskzRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_maskz_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRoundscaleRoundPd 

func M512MaskzRoundscaleRoundPd(k x86.Mmask8, a x86.M512d, imm8 byte, rounding int) (dst x86.M512d)

M512MaskzRoundscaleRoundPd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_maskz_roundscale_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRoundscaleRoundPs 

func M512MaskzRoundscaleRoundPs(k x86.Mmask16, a x86.M512, imm8 byte, rounding int) (dst x86.M512)

M512MaskzRoundscaleRoundPs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_maskz_roundscale_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzRsqrt14Pd 

func M512MaskzRsqrt14Pd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm512_maskz_rsqrt14_pd'. Requires AVX512F.

func M512MaskzRsqrt14Ps 

func M512MaskzRsqrt14Ps(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm512_maskz_rsqrt14_ps'. Requires AVX512F.

func M512MaskzScalefPd 

func M512MaskzScalefPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_maskz_scalef_pd'. Requires AVX512F.

func M512MaskzScalefPs 

func M512MaskzScalefPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_maskz_scalef_ps'. Requires AVX512F.

func M512MaskzScalefRoundPd 

func M512MaskzScalefRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskzScalefRoundPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_maskz_scalef_round_pd'. Requires AVX512F.

func M512MaskzScalefRoundPs 

func M512MaskzScalefRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskzScalefRoundPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_maskz_scalef_round_ps'. Requires AVX512F.

func M512MaskzSet1Epi32 

func M512MaskzSet1Epi32(k x86.Mmask16, a int) (dst x86.M512i)

M512MaskzSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_maskz_set1_epi32'. Requires AVX512F.

func M512MaskzSet1Epi64 

func M512MaskzSet1Epi64(k x86.Mmask8, a int64) (dst x86.M512i)

M512MaskzSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_maskz_set1_epi64'. Requires AVX512F.

func M512MaskzShuffleEpi32 

func M512MaskzShuffleEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzShuffleEpi32: Shuffle 32-bit integers in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(a[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(a[255:128], imm8[7:6])
tmp_dst[287:256] := SELECT4(a[383:256], imm8[1:0])
tmp_dst[319:288] := SELECT4(a[383:256], imm8[3:2])
tmp_dst[351:320] := SELECT4(a[383:256], imm8[5:4])
tmp_dst[383:352] := SELECT4(a[383:256], imm8[7:6])
tmp_dst[415:384] := SELECT4(a[511:384], imm8[1:0])
tmp_dst[447:416] := SELECT4(a[511:384], imm8[3:2])
tmp_dst[479:448] := SELECT4(a[511:384], imm8[5:4])
tmp_dst[511:480] := SELECT4(a[511:384], imm8[7:6])
FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSHUFD'. Intrinsic: '_mm512_maskz_shuffle_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShuffleF32x4 

func M512MaskzShuffleF32x4(k x86.Mmask16, a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512MaskzShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm512_maskz_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShuffleF64x2 

func M512MaskzShuffleF64x2(k x86.Mmask8, a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskzShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm512_maskz_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShuffleI32x4 

func M512MaskzShuffleI32x4(k x86.Mmask16, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm512_maskz_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShuffleI64x2 

func M512MaskzShuffleI64x2(k x86.Mmask8, a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

tmp_dst[127:0] := SELECT4(a[511:0], imm8[1:0])
tmp_dst[255:128] := SELECT4(a[511:0], imm8[3:2])
tmp_dst[383:256] := SELECT4(b[511:0], imm8[5:4])
tmp_dst[511:384] := SELECT4(b[511:0], imm8[7:6])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm512_maskz_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShufflePd 

func M512MaskzShufflePd(k x86.Mmask8, a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512MaskzShufflePd: Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]
tmp_dst[191:128] := (imm8[2] == 0) ? a[191:128] : a[255:192]
tmp_dst[255:192] := (imm8[3] == 0) ? b[191:128] : b[255:192]
tmp_dst[319:256] := (imm8[4] == 0) ? a[319:256] : a[383:320]
tmp_dst[383:320] := (imm8[5] == 0) ? b[319:256] : b[383:320]
tmp_dst[447:384] := (imm8[6] == 0) ? a[447:384] : a[511:448]
tmp_dst[511:448] := (imm8[7] == 0) ? b[447:384] : b[511:448]

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm512_maskz_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzShufflePs 

func M512MaskzShufflePs(k x86.Mmask16, a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512MaskzShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])
tmp_dst[159:128] := SELECT4(a[255:128], imm8[1:0])
tmp_dst[191:160] := SELECT4(a[255:128], imm8[3:2])
tmp_dst[223:192] := SELECT4(b[255:128], imm8[5:4])
tmp_dst[255:224] := SELECT4(b[255:128], imm8[7:6])
tmp_dst[287:256] := SELECT4(a[383:256], imm8[1:0])
tmp_dst[319:288] := SELECT4(a[383:256], imm8[3:2])
tmp_dst[351:320] := SELECT4(b[383:256], imm8[5:4])
tmp_dst[383:352] := SELECT4(b[383:256], imm8[7:6])
tmp_dst[415:384] := SELECT4(a[511:384], imm8[1:0])
tmp_dst[447:416] := SELECT4(a[511:384], imm8[3:2])
tmp_dst[479:448] := SELECT4(b[511:384], imm8[5:4])
tmp_dst[511:480] := SELECT4(b[511:384], imm8[7:6])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm512_maskz_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSllEpi32 

func M512MaskzSllEpi32(k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm512_maskz_sll_epi32'. Requires AVX512F.

func M512MaskzSllEpi64 

func M512MaskzSllEpi64(k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_maskz_sll_epi64'. Requires AVX512F.

func M512MaskzSlliEpi32 

func M512MaskzSlliEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSlliEpi32: Shift packed 32-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm512_maskz_slli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSlliEpi64 

func M512MaskzSlliEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_maskz_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSllvEpi32 

func M512MaskzSllvEpi32(k x86.Mmask16, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSllvEpi32: Shift packed 32-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLVD'. Intrinsic: '_mm512_maskz_sllv_epi32'. Requires AVX512F.

func M512MaskzSllvEpi64 

func M512MaskzSllvEpi64(k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm512_maskz_sllv_epi64'. Requires AVX512F.

func M512MaskzSqrtPd 

func M512MaskzSqrtPd(k x86.Mmask8, a x86.M512d) (dst x86.M512d)

M512MaskzSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_maskz_sqrt_pd'. Requires AVX512F.

func M512MaskzSqrtPs 

func M512MaskzSqrtPs(k x86.Mmask16, a x86.M512) (dst x86.M512)

M512MaskzSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_maskz_sqrt_ps'. Requires AVX512F.

func M512MaskzSqrtRoundPd 

func M512MaskzSqrtRoundPd(k x86.Mmask8, a x86.M512d, rounding int) (dst x86.M512d)

M512MaskzSqrtRoundPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE.

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := SQRT(a[i+63:i])
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_maskz_sqrt_round_pd'. Requires AVX512F.

func M512MaskzSqrtRoundPs 

func M512MaskzSqrtRoundPs(k x86.Mmask16, a x86.M512, rounding int) (dst x86.M512)

M512MaskzSqrtRoundPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := SQRT(a[i+31:i])
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_maskz_sqrt_round_ps'. Requires AVX512F.

func M512MaskzSraEpi32 

func M512MaskzSraEpi32(k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm512_maskz_sra_epi32'. Requires AVX512F.

func M512MaskzSraEpi64 

func M512MaskzSraEpi64(k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_maskz_sra_epi64'. Requires AVX512F.

func M512MaskzSraiEpi32 

func M512MaskzSraiEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSraiEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm512_maskz_srai_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSraiEpi64 

func M512MaskzSraiEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_maskz_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSravEpi32 

func M512MaskzSravEpi32(k x86.Mmask16, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSravEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAVD'. Intrinsic: '_mm512_maskz_srav_epi32'. Requires AVX512F.

func M512MaskzSravEpi64 

func M512MaskzSravEpi64(k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm512_maskz_srav_epi64'. Requires AVX512F.

func M512MaskzSrlEpi32 

func M512MaskzSrlEpi32(k x86.Mmask16, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm512_maskz_srl_epi32'. Requires AVX512F.

func M512MaskzSrlEpi64 

func M512MaskzSrlEpi64(k x86.Mmask8, a x86.M512i, count x86.M128i) (dst x86.M512i)

M512MaskzSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_maskz_srl_epi64'. Requires AVX512F.

func M512MaskzSrliEpi32 

func M512MaskzSrliEpi32(k x86.Mmask16, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSrliEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm512_maskz_srli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSrliEpi64 

func M512MaskzSrliEpi64(k x86.Mmask8, a x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_maskz_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzSrlvEpi32 

func M512MaskzSrlvEpi32(k x86.Mmask16, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSrlvEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLVD'. Intrinsic: '_mm512_maskz_srlv_epi32'. Requires AVX512F.

func M512MaskzSrlvEpi64 

func M512MaskzSrlvEpi64(k x86.Mmask8, a x86.M512i, count x86.M512i) (dst x86.M512i)

M512MaskzSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm512_maskz_srlv_epi64'. Requires AVX512F.

func M512MaskzSubEpi32 

func M512MaskzSubEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzSubEpi32: Subtract packed 32-bit integers in 'b' from packed 32-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSUBD'. Intrinsic: '_mm512_maskz_sub_epi32'. Requires AVX512F.

func M512MaskzSubEpi64 

func M512MaskzSubEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm512_maskz_sub_epi64'. Requires AVX512F.

func M512MaskzSubPd 

func M512MaskzSubPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzSubPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm512_maskz_sub_pd'. Requires AVX512F.

func M512MaskzSubPs 

func M512MaskzSubPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzSubPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm512_maskz_sub_ps'. Requires AVX512F.

func M512MaskzSubRoundPd 

func M512MaskzSubRoundPd(k x86.Mmask8, a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512MaskzSubRoundPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := j*64
			IF k[j]
				dst[i+63:i] := a[i+63:i] - b[i+63:i]
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm512_maskz_sub_round_pd'. Requires AVX512F.

func M512MaskzSubRoundPs 

func M512MaskzSubRoundPs(k x86.Mmask16, a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512MaskzSubRoundPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 15
			i := j*32
			IF k[j]
				dst[i+31:i] := a[i+31:i] - b[i+31:i]
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm512_maskz_sub_round_ps'. Requires AVX512F.

func M512MaskzTernarylogicEpi32 

func M512MaskzTernarylogicEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i, c x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm512_maskz_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzTernarylogicEpi64 

func M512MaskzTernarylogicEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i, c x86.M512i, imm8 byte) (dst x86.M512i)

M512MaskzTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm512_maskz_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512MaskzUnpackhiEpi32 

func M512MaskzUnpackhiEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm512_maskz_unpackhi_epi32'. Requires AVX512F.

func M512MaskzUnpackhiEpi64 

func M512MaskzUnpackhiEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm512_maskz_unpackhi_epi64'. Requires AVX512F.

func M512MaskzUnpackhiPd 

func M512MaskzUnpackhiPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm512_maskz_unpackhi_pd'. Requires AVX512F.

func M512MaskzUnpackhiPs 

func M512MaskzUnpackhiPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm512_maskz_unpackhi_ps'. Requires AVX512F.

func M512MaskzUnpackloEpi32 

func M512MaskzUnpackloEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm512_maskz_unpacklo_epi32'. Requires AVX512F.

func M512MaskzUnpackloEpi64 

func M512MaskzUnpackloEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm512_maskz_unpacklo_epi64'. Requires AVX512F.

func M512MaskzUnpackloPd 

func M512MaskzUnpackloPd(k x86.Mmask8, a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaskzUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm512_maskz_unpacklo_pd'. Requires AVX512F.

func M512MaskzUnpackloPs 

func M512MaskzUnpackloPs(k x86.Mmask16, a x86.M512, b x86.M512) (dst x86.M512)

M512MaskzUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
tmp_dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
tmp_dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
tmp_dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm512_maskz_unpacklo_ps'. Requires AVX512F.

func M512MaskzXorEpi32 

func M512MaskzXorEpi32(k x86.Mmask16, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzXorEpi32: Compute the bitwise XOR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 15
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] XOR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPXORD'. Intrinsic: '_mm512_maskz_xor_epi32'. Requires AVX512F.

func M512MaskzXorEpi64 

func M512MaskzXorEpi64(k x86.Mmask8, a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaskzXorEpi64: Compute the bitwise XOR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 7
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] XOR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_maskz_xor_epi64'. Requires AVX512F.

func M512MaxEpi64 

func M512MaxEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm512_max_epi64'. Requires AVX512F.

func M512MaxEpu64 

func M512MaxEpu64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm512_max_epu64'. Requires AVX512F.

func M512MaxPd 

func M512MaxPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_max_pd'. Requires AVX512F.

func M512MaxPs 

func M512MaxPs(a x86.M512, b x86.M512) (dst x86.M512)

M512MaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_max_ps'. Requires AVX512F.

func M512MaxRoundPd 

func M512MaxRoundPd(a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MaxRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm512_max_round_pd'. Requires AVX512F.

func M512MaxRoundPs 

func M512MaxRoundPs(a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MaxRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm512_max_round_ps'. Requires AVX512F.

func M512MinEpi64 

func M512MinEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm512_min_epi64'. Requires AVX512F.

func M512MinEpu64 

func M512MinEpu64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm512_min_epu64'. Requires AVX512F.

func M512MinPd 

func M512MinPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512MinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_min_pd'. Requires AVX512F.

func M512MinPs 

func M512MinPs(a x86.M512, b x86.M512) (dst x86.M512)

M512MinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_min_ps'. Requires AVX512F.

func M512MinRoundPd 

func M512MinRoundPd(a x86.M512d, b x86.M512d, sae int) (dst x86.M512d)

M512MinRoundPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 7
		i := j*64
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm512_min_round_pd'. Requires AVX512F.

func M512MinRoundPs 

func M512MinRoundPs(a x86.M512, b x86.M512, sae int) (dst x86.M512)

M512MinRoundPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	FOR j := 0 to 15
		i := j*32
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ENDFOR
	dst[MAX:512] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm512_min_round_ps'. Requires AVX512F.

func M512MovedupPd 

func M512MovedupPd(a x86.M512d) (dst x86.M512d)

M512MovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst'. 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
tmp[191:128] := a[191:128]
tmp[255:192] := a[191:128]
tmp[319:256] := a[319:256]
tmp[383:320] := a[319:256]
tmp[447:384] := a[447:384]
tmp[511:448] := a[447:384]
dst[MAX:512] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm512_movedup_pd'. Requires AVX512F.

func M512MovehdupPs 

func M512MovehdupPs(a x86.M512) (dst x86.M512)

M512MovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst'. 

dst[31:0] := a[63:32]
dst[63:32] := a[63:32]
dst[95:64] := a[127:96]
dst[127:96] := a[127:96]
dst[159:128] := a[191:160]
dst[191:160] := a[191:160]
dst[223:192] := a[255:224]
dst[255:224] := a[255:224]
dst[287:256] := a[319:288]
dst[319:288] := a[319:288]
dst[351:320] := a[383:352]
dst[383:352] := a[383:352]
dst[415:384] := a[447:416]
dst[447:416] := a[447:416]
dst[479:448] := a[511:480]
dst[511:480] := a[511:480]
dst[MAX:512] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm512_movehdup_ps'. Requires AVX512F.

func M512MoveldupPs 

func M512MoveldupPs(a x86.M512) (dst x86.M512)

M512MoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst'. 

dst[31:0] := a[31:0]
dst[63:32] := a[31:0]
dst[95:64] := a[95:64]
dst[127:96] := a[95:64]
dst[159:128] := a[159:128]
dst[191:160] := a[159:128]
dst[223:192] := a[223:192]
dst[255:224] := a[223:192]
dst[287:256] := a[287:256]
dst[319:288] := a[287:256]
dst[351:320] := a[351:320]
dst[383:352] := a[351:320]
dst[415:384] := a[415:384]
dst[447:416] := a[415:384]
dst[479:448] := a[479:448]
dst[511:480] := a[479:448]
dst[MAX:512] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm512_moveldup_ps'. Requires AVX512F.

func M512MulEpi32 

func M512MulEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[i+31:i] * b[i+31:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm512_mul_epi32'. Requires AVX512F.

func M512MulEpu32 

func M512MulEpu32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[i+31:i] * b[i+31:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm512_mul_epu32'. Requires AVX512F.

func M512MulloxEpi64 

func M512MulloxEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512MulloxEpi64: Multiplies elements in packed 64-bit integer vectors 'a' and 'b' together, storing the lower 64 bits of the result in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[i+63:i] * b[i+63:i]
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_mullox_epi64'. Requires AVX512F.

func M512NearbyintPd 

func M512NearbyintPd(a x86.M512d) (dst x86.M512d)

M512NearbyintPd: Rounds each packed double-precision (64-bit) floating-point element in 'a' to the nearest integer value and stores the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := NearbyInt(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_nearbyint_pd'. Requires AVX512F.

func M512NearbyintPs 

func M512NearbyintPs(a x86.M512) (dst x86.M512)

M512NearbyintPs: Rounds each packed single-precision (32-bit) floating-point element in 'a' to the nearest integer value and stores the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := NearbyInt(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_nearbyint_ps'. Requires AVX512F.

func M512PermutePd 

func M512PermutePd(a x86.M512d, imm8 byte) (dst x86.M512d)

M512PermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst'. 

IF (imm8[0] == 0) dst[63:0] := a[63:0]
IF (imm8[0] == 1) dst[63:0] := a[127:64]
IF (imm8[1] == 0) dst[127:64] := a[63:0]
IF (imm8[1] == 1) dst[127:64] := a[127:64]
IF (imm8[2] == 0) dst[191:128] := a[191:128]
IF (imm8[2] == 1) dst[191:128] := a[255:192]
IF (imm8[3] == 0) dst[255:192] := a[191:128]
IF (imm8[3] == 1) dst[255:192] := a[255:192]
IF (imm8[4] == 0) dst[319:256] := a[319:256]
IF (imm8[4] == 1) dst[319:256] := a[383:320]
IF (imm8[5] == 0) dst[383:320] := a[319:256]
IF (imm8[5] == 1) dst[383:320] := a[383:320]
IF (imm8[6] == 0) dst[447:384] := a[447:384]
IF (imm8[6] == 1) dst[447:384] := a[511:448]
IF (imm8[7] == 0) dst[511:448] := a[447:384]
IF (imm8[7] == 1) dst[511:448] := a[511:448]
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512PermutePs 

func M512PermutePs(a x86.M512, imm8 byte) (dst x86.M512)

M512PermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

dst[31:0] := SELECT4(a[127:0], imm8[1:0])
dst[63:32] := SELECT4(a[127:0], imm8[3:2])
dst[95:64] := SELECT4(a[127:0], imm8[5:4])
dst[127:96] := SELECT4(a[127:0], imm8[7:6])
dst[159:128] := SELECT4(a[255:128], imm8[1:0])
dst[191:160] := SELECT4(a[255:128], imm8[3:2])
dst[223:192] := SELECT4(a[255:128], imm8[5:4])
dst[255:224] := SELECT4(a[255:128], imm8[7:6])
dst[287:256] := SELECT4(a[383:256], imm8[1:0])
dst[319:288] := SELECT4(a[383:256], imm8[3:2])
dst[351:320] := SELECT4(a[383:256], imm8[5:4])
dst[383:352] := SELECT4(a[383:256], imm8[7:6])
dst[415:384] := SELECT4(a[511:384], imm8[1:0])
dst[447:416] := SELECT4(a[511:384], imm8[3:2])
dst[479:448] := SELECT4(a[511:384], imm8[5:4])
dst[511:480] := SELECT4(a[511:384], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512PermutevarPd 

func M512PermutevarPd(a x86.M512d, b x86.M512i) (dst x86.M512d)

M512PermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst'. 

IF (b[1] == 0) dst[63:0] := a[63:0]
IF (b[1] == 1) dst[63:0] := a[127:64]
IF (b[65] == 0) dst[127:64] := a[63:0]
IF (b[65] == 1) dst[127:64] := a[127:64]
IF (b[129] == 0) dst[191:128] := a[191:128]
IF (b[129] == 1) dst[191:128] := a[255:192]
IF (b[193] == 0) dst[255:192] := a[191:128]
IF (b[193] == 1) dst[255:192] := a[255:192]
IF (b[257] == 0) dst[319:256] := a[319:256]
IF (b[257] == 1) dst[319:256] := a[383:320]
IF (b[321] == 0) dst[383:320] := a[319:256]
IF (b[321] == 1) dst[383:320] := a[383:320]
IF (b[385] == 0) dst[447:384] := a[447:384]
IF (b[385] == 1) dst[447:384] := a[511:448]
IF (b[449] == 0) dst[511:448] := a[447:384]
IF (b[449] == 1) dst[511:448] := a[511:448]
dst[MAX:512] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm512_permutevar_pd'. Requires AVX512F.

func M512PermutevarPs 

func M512PermutevarPs(a x86.M512, b x86.M512i) (dst x86.M512)

M512PermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

dst[31:0] := SELECT4(a[127:0], b[1:0])
dst[63:32] := SELECT4(a[127:0], b[33:32])
dst[95:64] := SELECT4(a[127:0], b[65:64])
dst[127:96] := SELECT4(a[127:0], b[97:96])
dst[159:128] := SELECT4(a[255:128], b[129:128])
dst[191:160] := SELECT4(a[255:128], b[161:160])
dst[223:192] := SELECT4(a[255:128], b[193:192])
dst[255:224] := SELECT4(a[255:128], b[225:224])
dst[287:256] := SELECT4(a[383:256], b[257:256])
dst[319:288] := SELECT4(a[383:256], b[289:288])
dst[351:320] := SELECT4(a[383:256], b[321:320])
dst[383:352] := SELECT4(a[383:256], b[353:352])
dst[415:384] := SELECT4(a[511:384], b[385:384])
dst[447:416] := SELECT4(a[511:384], b[417:416])
dst[479:448] := SELECT4(a[511:384], b[449:448])
dst[511:480] := SELECT4(a[511:384], b[481:480])
dst[MAX:512] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm512_permutevar_ps'. Requires AVX512F.

func M512Permutex2varEpi32 

func M512Permutex2varEpi32(a x86.M512i, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm512_permutex2var_epi32'. Requires AVX512F.

func M512Permutex2varEpi64 

func M512Permutex2varEpi64(a x86.M512i, idx x86.M512i, b x86.M512i) (dst x86.M512i)

M512Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm512_permutex2var_epi64'. Requires AVX512F.

func M512Permutex2varPd 

func M512Permutex2varPd(a x86.M512d, idx x86.M512i, b x86.M512d) (dst x86.M512d)

M512Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	off := idx[i+2:i]*64
	dst[i+63:i] := idx[i+3] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm512_permutex2var_pd'. Requires AVX512F.

func M512Permutex2varPs 

func M512Permutex2varPs(a x86.M512, idx x86.M512i, b x86.M512) (dst x86.M512)

M512Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' across lanes using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	off := idx[i+3:i]*32
	dst[i+31:i] := idx[i+4] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm512_permutex2var_ps'. Requires AVX512F.

func M512PermutexEpi64 

func M512PermutexEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512PermutexEpi64: Shuffle 64-bit integers in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

dst[63:0] := SELECT4(a[255:0], imm8[1:0])
dst[127:64] := SELECT4(a[255:0], imm8[3:2])
dst[191:128] := SELECT4(a[255:0], imm8[5:4])
dst[255:192] := SELECT4(a[255:0], imm8[7:6])
dst[319:256] := SELECT4(a[511:256], imm8[1:0])
dst[383:320] := SELECT4(a[511:256], imm8[3:2])
dst[447:384] := SELECT4(a[511:256], imm8[5:4])
dst[511:448] := SELECT4(a[511:256], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_permutex_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512PermutexPd 

func M512PermutexPd(a x86.M512d, imm8 byte) (dst x86.M512d)

M512PermutexPd: Shuffle double-precision (64-bit) floating-point elements in 'a' within 256-bit lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[63:0] := src[63:0]
	1:	tmp[63:0] := src[127:64]
	2:	tmp[63:0] := src[191:128]
	3:	tmp[63:0] := src[255:192]
	ESAC
	RETURN tmp[63:0]
}

dst[63:0] := SELECT4(a[255:0], imm8[1:0])
dst[127:64] := SELECT4(a[255:0], imm8[3:2])
dst[191:128] := SELECT4(a[255:0], imm8[5:4])
dst[255:192] := SELECT4(a[255:0], imm8[7:6])
dst[319:256] := SELECT4(a[511:256], imm8[1:0])
dst[383:320] := SELECT4(a[511:256], imm8[3:2])
dst[447:384] := SELECT4(a[511:256], imm8[5:4])
dst[511:448] := SELECT4(a[511:256], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_permutex_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512PermutexvarEpi32 

func M512PermutexvarEpi32(idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512PermutexvarEpi32: Shuffle 32-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	dst[i+31:i] := a[id+31:id]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMD'. Intrinsic: '_mm512_permutexvar_epi32'. Requires AVX512F.

func M512PermutexvarEpi64 

func M512PermutexvarEpi64(idx x86.M512i, a x86.M512i) (dst x86.M512i)

M512PermutexvarEpi64: Shuffle 64-bit integers in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	dst[i+63:i] := a[id+63:id]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMQ'. Intrinsic: '_mm512_permutexvar_epi64'. Requires AVX512F.

func M512PermutexvarPd 

func M512PermutexvarPd(idx x86.M512i, a x86.M512d) (dst x86.M512d)

M512PermutexvarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	id := idx[i+2:i]*64
	dst[i+63:i] := a[id+63:id]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPD'. Intrinsic: '_mm512_permutexvar_pd'. Requires AVX512F.

func M512PermutexvarPs 

func M512PermutexvarPs(idx x86.M512i, a x86.M512) (dst x86.M512)

M512PermutexvarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' across lanes using the corresponding index in 'idx'. 

FOR j := 0 to 15
	i := j*32
	id := idx[i+3:i]*32
	dst[i+31:i] := a[id+31:id]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPERMPS'. Intrinsic: '_mm512_permutexvar_ps'. Requires AVX512F.

func M512PowPd 

func M512PowPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512PowPd: Compute the exponential value of packed double-precision (64-bit) floating-point elements in 'a' raised by packed elements in 'b', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := (a[i+63:i])^(b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_pow_pd'. Requires AVX512F.

func M512PowPs 

func M512PowPs(a x86.M512, b x86.M512) (dst x86.M512)

M512PowPs: Compute the exponential value of packed single-precision (32-bit) floating-point elements in 'a' raised by packed elements in 'b', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := (a[i+31:i])^(b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_pow_ps'. Requires AVX512F.

func M512Rcp14Pd 

func M512Rcp14Pd(a x86.M512d) (dst x86.M512d)

M512Rcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm512_rcp14_pd'. Requires AVX512F.

func M512Rcp14Ps 

func M512Rcp14Ps(a x86.M512) (dst x86.M512)

M512Rcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm512_rcp14_ps'. Requires AVX512F.

func M512RecipPd 

func M512RecipPd(a x86.M512d) (dst x86.M512d)

M512RecipPd: Computes the reciprocal of packed double-precision (64-bit) floating-point elements in 'a', storing the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := (1 / a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_recip_pd'. Requires AVX512F.

func M512RecipPs 

func M512RecipPs(a x86.M512) (dst x86.M512)

M512RecipPs: Computes the reciprocal of packed single-precision (32-bit) floating-point elements in 'a', storing the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := (1 / a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_recip_ps'. Requires AVX512F.

func M512RemEpi16 

func M512RemEpi16(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpi16: Divide packed 16-bit integers in 'a' by packed elements in 'b', and store the remainders as packed 32-bit integers in 'dst'. 

FOR j := 0 to 31
	i := 16*j
	dst[i+15:i] := REMAINDER(a[i+15:i] / b[i+15:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epi16'. Requires AVX512F.

func M512RemEpi32 

func M512RemEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpi32: Divide packed 32-bit integers in 'a' by packed elements in 'b', and store the remainders as packed 32-bit integers in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := REMAINDER(a[i+31:i] / b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epi32'. Requires AVX512F.

func M512RemEpi64 

func M512RemEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpi64: Divide packed 64-bit integers in 'a' by packed elements in 'b', and store the remainders as packed 32-bit integers in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	dst[i+63:i] := REMAINDER(a[i+63:i] / b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epi64'. Requires AVX512F.

func M512RemEpi8 

func M512RemEpi8(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpi8: Divide packed 8-bit integers in 'a' by packed elements in 'b', and store the remainders as packed 32-bit integers in 'dst'. 

FOR j := 0 to 63
	i := 8*j
	dst[i+7:i] := REMAINDER(a[i+7:i] / b[i+7:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epi8'. Requires AVX512F.

func M512RemEpu16 

func M512RemEpu16(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpu16: Divide packed unsigned 16-bit integers in 'a' by packed elements in 'b', and store the remainders as packed unsigned 32-bit integers in 'dst'. 

FOR j := 0 to 31
	i := 16*j
	dst[i+15:i] := REMAINDER(a[i+15:i] / b[i+15:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epu16'. Requires AVX512F.

func M512RemEpu32 

func M512RemEpu32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpu32: Divide packed unsigned 32-bit integers in 'a' by packed elements in 'b', and store the remainders as packed unsigned 32-bit integers in 'dst'. 

FOR j := 0 to 15
	i := 32*j
	dst[i+31:i] := REMAINDER(a[i+31:i] / b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epu32'. Requires AVX512F.

func M512RemEpu64 

func M512RemEpu64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpu64: Divide packed unsigned 64-bit integers in 'a' by packed elements in 'b', and store the remainders as packed unsigned 32-bit integers in 'dst'. 

FOR j := 0 to 7
	i := 64*j
	dst[i+63:i] := REMAINDER(a[i+63:i] / b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epu64'. Requires AVX512F.

func M512RemEpu8 

func M512RemEpu8(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RemEpu8: Divide packed unsigned 8-bit integers in 'a' by packed elements in 'b', and store the remainders as packed unsigned 32-bit integers in 'dst'. 

FOR j := 0 to 63
	i := 8*j
	dst[i+7:i] := REMAINDER(a[i+7:i] / b[i+7:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rem_epu8'. Requires AVX512F.

func M512RintPd 

func M512RintPd(a x86.M512d) (dst x86.M512d)

M512RintPd: Rounds the packed double-precision (64-bit) floating-point elements in 'a' to the nearest even integer value and stores the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := RoundToNearestEven(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rint_pd'. Requires AVX512F.

func M512RintPs 

func M512RintPs(a x86.M512) (dst x86.M512)

M512RintPs: Rounds the packed single-precision (32-bit) floating-point elements in 'a' to the nearest even integer value and stores the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := RoundToNearestEven(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_rint_ps'. Requires AVX512F.

func M512RolEpi32 

func M512RolEpi32(a x86.M512i, imm8 byte) (dst x86.M512i)

M512RolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm512_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RolEpi64 

func M512RolEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512RolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm512_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RolvEpi32 

func M512RolvEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm512_rolv_epi32'. Requires AVX512F.

func M512RolvEpi64 

func M512RolvEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm512_rolv_epi64'. Requires AVX512F.

func M512RorEpi32 

func M512RorEpi32(a x86.M512i, imm8 byte) (dst x86.M512i)

M512RorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm512_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RorEpi64 

func M512RorEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512RorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm512_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RorvEpi32 

func M512RorvEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm512_rorv_epi32'. Requires AVX512F.

func M512RorvEpi64 

func M512RorvEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512RorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm512_rorv_epi64'. Requires AVX512F.

func M512RoundscalePd 

func M512RoundscalePd(a x86.M512d, imm8 byte) (dst x86.M512d)

M512RoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RoundscalePs 

func M512RoundscalePs(a x86.M512, imm8 byte) (dst x86.M512)

M512RoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RoundscaleRoundPd 

func M512RoundscaleRoundPd(a x86.M512d, imm8 byte, rounding int) (dst x86.M512d)

M512RoundscaleRoundPd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm512_roundscale_round_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512RoundscaleRoundPs 

func M512RoundscaleRoundPs(a x86.M512, imm8 byte, rounding int) (dst x86.M512)

M512RoundscaleRoundPs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm512_roundscale_round_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512Rsqrt14Pd 

func M512Rsqrt14Pd(a x86.M512d) (dst x86.M512d)

M512Rsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm512_rsqrt14_pd'. Requires AVX512F.

func M512Rsqrt14Ps 

func M512Rsqrt14Ps(a x86.M512) (dst x86.M512)

M512Rsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
ENDFOR
dst[MAX:512] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm512_rsqrt14_ps'. Requires AVX512F.

func M512ScalefPd 

func M512ScalefPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512ScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_scalef_pd'. Requires AVX512F.

func M512ScalefPs 

func M512ScalefPs(a x86.M512, b x86.M512) (dst x86.M512)

M512ScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_scalef_ps'. Requires AVX512F.

func M512ScalefRoundPd 

func M512ScalefRoundPd(a x86.M512d, b x86.M512d, rounding int) (dst x86.M512d)

M512ScalefRoundPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		FOR j := 0 to 7
			i := j*64
			dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm512_scalef_round_pd'. Requires AVX512F.

func M512ScalefRoundPs 

func M512ScalefRoundPs(a x86.M512, b x86.M512, rounding int) (dst x86.M512)

M512ScalefRoundPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[31:0]
		}

		FOR j := 0 to 15
			i := j*32
			dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm512_scalef_round_ps'. Requires AVX512F.

func M512Set1Epi16 

func M512Set1Epi16(a int16) (dst x86.M512i)

M512Set1Epi16: Broadcast the low packed 16-bit integer from 'a' to all all elements of 'dst'. 

FOR j := 0 to 31
	i := j*16
	dst[i+15:i] := a[15:0]
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set1_epi16'. Requires AVX512F.

func M512Set1Epi32 

func M512Set1Epi32(a int) (dst x86.M512i)

M512Set1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := a[31:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm512_set1_epi32'. Requires AVX512F.

func M512Set1Epi64 

func M512Set1Epi64(a int64) (dst x86.M512i)

M512Set1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[63:0]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm512_set1_epi64'. Requires AVX512F.

func M512Set1Epi8 

func M512Set1Epi8(a byte) (dst x86.M512i)

M512Set1Epi8: Broadcast 8-bit integer 'a' to all elements of 'dst'. 

FOR j := 0 to 63
	i := j*8
	dst[i+7:i] := a[7:0]
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set1_epi8'. Requires AVX512F.

func M512Set1Pd 

func M512Set1Pd(a float64) (dst x86.M512d)

M512Set1Pd: Broadcast double-precision (64-bit) floating-point value 'a' to all elements of 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[63:0]
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set1_pd'. Requires AVX512F.

func M512Set1Ps 

func M512Set1Ps(a float32) (dst x86.M512)

M512Set1Ps: Broadcast single-precision (32-bit) floating-point value 'a' to all elements of 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := a[31:0]
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set1_ps'. Requires AVX512F.

func M512Set4Epi32 

func M512Set4Epi32(d int, c int, b int, a int) (dst x86.M512i)

M512Set4Epi32: Set packed 32-bit integers in 'dst' with the repeated 4 element sequence. 

dst[31:0] := d
dst[63:32] := c
dst[95:64] := b
dst[127:96] := a
dst[159:128] := d
dst[191:160] := c
dst[223:192] := b
dst[255:224] := a
dst[287:256] := d
dst[319:288] := c
dst[351:320] := b
dst[383:352] := a
dst[415:384] := d
dst[447:416] := c
dst[479:448] := b
dst[511:480] := a
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set4_epi32'. Requires AVX512F.

func M512Set4Epi64 

func M512Set4Epi64(d int64, c int64, b int64, a int64) (dst x86.M512i)

M512Set4Epi64: Set packed 64-bit integers in 'dst' with the repeated 4 element sequence. 

dst[63:0] := d
dst[127:64] := c
dst[191:128] := b
dst[255:192] := a
dst[319:256] := d
dst[383:320] := c
dst[447:384] := b
dst[511:448] := a
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set4_epi64'. Requires AVX512F.

func M512Set4Pd 

func M512Set4Pd(d float64, c float64, b float64, a float64) (dst x86.M512d)

M512Set4Pd: Set packed double-precision (64-bit) floating-point elements in 'dst' with the repeated 4 element sequence. 

dst[63:0] := d
dst[127:64] := c
dst[191:128] := b
dst[255:192] := a
dst[319:256] := d
dst[383:320] := c
dst[447:384] := b
dst[511:448] := a
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set4_pd'. Requires AVX512F.

func M512Set4Ps 

func M512Set4Ps(d float32, c float32, b float32, a float32) (dst x86.M512)

M512Set4Ps: Set packed single-precision (32-bit) floating-point elements in 'dst' with the repeated 4 element sequence. 

dst[31:0] := d
dst[63:32] := c
dst[95:64] := b
dst[127:96] := a
dst[159:128] := d
dst[191:160] := c
dst[223:192] := b
dst[255:224] := a
dst[287:256] := d
dst[319:288] := c
dst[351:320] := b
dst[383:352] := a
dst[415:384] := d
dst[447:416] := c
dst[479:448] := b
dst[511:480] := a
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set4_ps'. Requires AVX512F.

func M512SetEpi32 

func M512SetEpi32(e15 int, e14 int, e13 int, e12 int, e11 int, e10 int, e9 int, e8 int, e7 int, e6 int, e5 int, e4 int, e3 int, e2 int, e1 int, e0 int) (dst x86.M512i)

M512SetEpi32: Set packed 32-bit integers in 'dst' with the supplied values. 

dst[31:0] := e0
dst[63:32] := e1
dst[95:64] := e2
dst[127:96] := e3
dst[159:128] := e4
dst[191:160] := e5
dst[223:192] := e6
dst[255:224] := e7
dst[287:256] := e8
dst[319:288] := e9
dst[351:320] := e10
dst[383:352] := e11
dst[415:384] := e12
dst[447:416] := e13
dst[479:448] := e14
dst[511:480] := e15
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set_epi32'. Requires AVX512F.

func M512SetEpi64 

func M512SetEpi64(e7 int64, e6 int64, e5 int64, e4 int64, e3 int64, e2 int64, e1 int64, e0 int64) (dst x86.M512i)

M512SetEpi64: Set packed 64-bit integers in 'dst' with the supplied values. 

dst[63:0] := e0
dst[127:64] := e1
dst[191:128] := e2
dst[255:192] := e3
dst[319:256] := e4
dst[383:320] := e5
dst[447:384] := e6
dst[511:448] := e7
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set_epi64'. Requires AVX512F.

func M512SetPd 

func M512SetPd(e7 float64, e6 float64, e5 float64, e4 float64, e3 float64, e2 float64, e1 float64, e0 float64) (dst x86.M512d)

M512SetPd: Set packed double-precision (64-bit) floating-point elements in 'dst' with the supplied values. 

dst[63:0] := e0
dst[127:64] := e1
dst[191:128] := e2
dst[255:192] := e3
dst[319:256] := e4
dst[383:320] := e5
dst[447:384] := e6
dst[511:448] := e7
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set_pd'. Requires AVX512F.

func M512SetPs 

func M512SetPs(e15 float32, e14 float32, e13 float32, e12 float32, e11 float32, e10 float32, e9 float32, e8 float32, e7 float32, e6 float32, e5 float32, e4 float32, e3 float32, e2 float32, e1 float32, e0 float32) (dst x86.M512)

M512SetPs: Set packed single-precision (32-bit) floating-point elements in 'dst' with the supplied values. 

dst[31:0] := e0
dst[63:32] := e1
dst[95:64] := e2
dst[127:96] := e3
dst[159:128] := e4
dst[191:160] := e5
dst[223:192] := e6
dst[255:224] := e7
dst[287:256] := e8
dst[319:288] := e9
dst[351:320] := e10
dst[383:352] := e11
dst[415:384] := e12
dst[447:416] := e13
dst[479:448] := e14
dst[511:480] := e15
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_set_ps'. Requires AVX512F.

func M512Setr4Epi32 

func M512Setr4Epi32(d int, c int, b int, a int) (dst x86.M512i)

M512Setr4Epi32: Set packed 32-bit integers in 'dst' with the repeated 4 element sequence in reverse order. 

dst[31:0] := a
dst[63:32] := b
dst[95:64] := c
dst[127:96] := d
dst[159:128] := a
dst[191:160] := b
dst[223:192] := c
dst[255:224] := d
dst[287:256] := a
dst[319:288] := b
dst[351:320] := c
dst[383:352] := d
dst[415:384] := a
dst[447:416] := b
dst[479:448] := c
dst[511:480] := d
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr4_epi32'. Requires AVX512F.

func M512Setr4Epi64 

func M512Setr4Epi64(d int64, c int64, b int64, a int64) (dst x86.M512i)

M512Setr4Epi64: Set packed 64-bit integers in 'dst' with the repeated 4 element sequence in reverse order. 

dst[63:0] := a
dst[127:64] := b
dst[191:128] := c
dst[255:192] := d
dst[319:256] := a
dst[383:320] := b
dst[447:384] := c
dst[511:448] := d
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr4_epi64'. Requires AVX512F.

func M512Setr4Pd 

func M512Setr4Pd(d float64, c float64, b float64, a float64) (dst x86.M512d)

M512Setr4Pd: Set packed double-precision (64-bit) floating-point elements in 'dst' with the repeated 4 element sequence in reverse order. 

dst[63:0] := a
dst[127:64] := b
dst[191:128] := c
dst[255:192] := d
dst[319:256] := a
dst[383:320] := b
dst[447:384] := c
dst[511:448] := d
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr4_pd'. Requires AVX512F.

func M512Setr4Ps 

func M512Setr4Ps(d float32, c float32, b float32, a float32) (dst x86.M512)

M512Setr4Ps: Set packed single-precision (32-bit) floating-point elements in 'dst' with the repeated 4 element sequence in reverse order. 

dst[31:0] := a
dst[63:32] := b
dst[95:64] := c
dst[127:96] := d
dst[159:128] := a
dst[191:160] := b
dst[223:192] := c
dst[255:224] := d
dst[287:256] := a
dst[319:288] := b
dst[351:320] := c
dst[383:352] := d
dst[415:384] := a
dst[447:416] := b
dst[479:448] := c
dst[511:480] := d
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr4_ps'. Requires AVX512F.

func M512SetrEpi32 

func M512SetrEpi32(e15 int, e14 int, e13 int, e12 int, e11 int, e10 int, e9 int, e8 int, e7 int, e6 int, e5 int, e4 int, e3 int, e2 int, e1 int, e0 int) (dst x86.M512i)

M512SetrEpi32: Set packed 32-bit integers in 'dst' with the supplied values in reverse order. 

dst[31:0] := e15
dst[63:32] := e14
dst[95:64] := e13
dst[127:96] := e12
dst[159:128] := e11
dst[191:160] := e10
dst[223:192] := e9
dst[255:224] := e8
dst[287:256] := e7
dst[319:288] := e6
dst[351:320] := e5
dst[383:352] := e4
dst[415:384] := e3
dst[447:416] := e2
dst[479:448] := e1
dst[511:480] := e0
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr_epi32'. Requires AVX512F.

func M512SetrEpi64 

func M512SetrEpi64(e7 int64, e6 int64, e5 int64, e4 int64, e3 int64, e2 int64, e1 int64, e0 int64) (dst x86.M512i)

M512SetrEpi64: Set packed 64-bit integers in 'dst' with the supplied values in reverse order. 

dst[63:0] := e7
dst[127:64] := e6
dst[191:128] := e5
dst[255:192] := e4
dst[319:256] := e3
dst[383:320] := e2
dst[447:384] := e1
dst[511:448] := e0
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr_epi64'. Requires AVX512F.

func M512SetrPd 

func M512SetrPd(e7 float64, e6 float64, e5 float64, e4 float64, e3 float64, e2 float64, e1 float64, e0 float64) (dst x86.M512d)

M512SetrPd: Set packed double-precision (64-bit) floating-point elements in 'dst' with the supplied values in reverse order. 

dst[63:0] := e7
dst[127:64] := e6
dst[191:128] := e5
dst[255:192] := e4
dst[319:256] := e3
dst[383:320] := e2
dst[447:384] := e1
dst[511:448] := e0
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr_pd'. Requires AVX512F.

func M512SetrPs 

func M512SetrPs(e15 float32, e14 float32, e13 float32, e12 float32, e11 float32, e10 float32, e9 float32, e8 float32, e7 float32, e6 float32, e5 float32, e4 float32, e3 float32, e2 float32, e1 float32, e0 float32) (dst x86.M512)

M512SetrPs: Set packed single-precision (32-bit) floating-point elements in 'dst' with the supplied values in reverse order. 

dst[31:0] := e15
dst[63:32] := e14
dst[95:64] := e13
dst[127:96] := e12
dst[159:128] := e11
dst[191:160] := e10
dst[223:192] := e9
dst[255:224] := e8
dst[287:256] := e7
dst[319:288] := e6
dst[351:320] := e5
dst[383:352] := e4
dst[415:384] := e3
dst[447:416] := e2
dst[479:448] := e1
dst[511:480] := e0
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_setr_ps'. Requires AVX512F.

func M512Setzero 

func M512Setzero() (dst x86.M512)

M512Setzero: Return vector of type __m512 with all elements set to zero. 

dst[MAX:0] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_setzero'. Requires AVX512F.

func M512SetzeroEpi32 

func M512SetzeroEpi32() (dst x86.M512i)

M512SetzeroEpi32: Return vector of type __m512i with all elements set to zero. 

dst[MAX:0] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_setzero_epi32'. Requires AVX512F.

func M512SetzeroPd 

func M512SetzeroPd() (dst x86.M512d)

M512SetzeroPd: Return vector of type __m512d with all elements set to zero. 

dst[MAX:0] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_setzero_pd'. Requires AVX512F.

func M512SetzeroPs 

func M512SetzeroPs() (dst x86.M512)

M512SetzeroPs: Return vector of type __m512 with all elements set to zero. 

dst[MAX:0] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_setzero_ps'. Requires AVX512F.

func M512SetzeroSi512 

func M512SetzeroSi512() (dst x86.M512i)

M512SetzeroSi512: Return vector of type __m512i with all elements set to zero. 

dst[MAX:0] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm512_setzero_si512'. Requires AVX512F.

func M512ShuffleF32x4 

func M512ShuffleF32x4(a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512ShuffleF32x4: Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT4(a[511:0], imm8[1:0])
dst[255:128] := SELECT4(a[511:0], imm8[3:2])
dst[383:256] := SELECT4(b[511:0], imm8[5:4])
dst[511:384] := SELECT4(b[511:0], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VSHUFF32X4'. Intrinsic: '_mm512_shuffle_f32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512ShuffleF64x2 

func M512ShuffleF64x2(a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512ShuffleF64x2: Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT4(a[511:0], imm8[1:0])
dst[255:128] := SELECT4(a[511:0], imm8[3:2])
dst[383:256] := SELECT4(b[511:0], imm8[5:4])
dst[511:384] := SELECT4(b[511:0], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VSHUFF64X2'. Intrinsic: '_mm512_shuffle_f64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512ShuffleI32x4 

func M512ShuffleI32x4(a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512ShuffleI32x4: Shuffle 128-bits (composed of 4 32-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT4(a[511:0], imm8[1:0])
dst[255:128] := SELECT4(a[511:0], imm8[3:2])
dst[383:256] := SELECT4(b[511:0], imm8[5:4])
dst[511:384] := SELECT4(b[511:0], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VSHUFI32X4'. Intrinsic: '_mm512_shuffle_i32x4'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512ShuffleI64x2 

func M512ShuffleI64x2(a x86.M512i, b x86.M512i, imm8 byte) (dst x86.M512i)

M512ShuffleI64x2: Shuffle 128-bits (composed of 2 64-bit integers) selected by 'imm8' from 'a' and 'b', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[127:0] := src[127:0]
	1:	tmp[127:0] := src[255:128]
	2:	tmp[127:0] := src[383:256]
	3:	tmp[127:0] := src[511:384]
	ESAC
	RETURN tmp[127:0]
}

dst[127:0] := SELECT4(a[511:0], imm8[1:0])
dst[255:128] := SELECT4(a[511:0], imm8[3:2])
dst[383:256] := SELECT4(b[511:0], imm8[5:4])
dst[511:384] := SELECT4(b[511:0], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VSHUFI64X2'. Intrinsic: '_mm512_shuffle_i64x2'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512ShufflePd 

func M512ShufflePd(a x86.M512d, b x86.M512d, imm8 byte) (dst x86.M512d)

M512ShufflePd: Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in 'imm8', and store the results in 'dst'. 

dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]
dst[191:128] := (imm8[2] == 0) ? a[191:128] : a[255:192]
dst[255:192] := (imm8[3] == 0) ? b[191:128] : b[255:192]
dst[319:256] := (imm8[4] == 0) ? a[319:256] : a[383:320]
dst[383:320] := (imm8[5] == 0) ? b[319:256] : b[383:320]
dst[447:384] := (imm8[6] == 0) ? a[447:384] : a[511:448]
dst[511:448] := (imm8[7] == 0) ? b[447:384] : b[511:448]
dst[MAX:512] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm512_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512ShufflePs 

func M512ShufflePs(a x86.M512, b x86.M512, imm8 byte) (dst x86.M512)

M512ShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' within 128-bit lanes using the control in 'imm8', and store the results in 'dst'. 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

dst[31:0] := SELECT4(a[127:0], imm8[1:0])
dst[63:32] := SELECT4(a[127:0], imm8[3:2])
dst[95:64] := SELECT4(b[127:0], imm8[5:4])
dst[127:96] := SELECT4(b[127:0], imm8[7:6])
dst[159:128] := SELECT4(a[255:128], imm8[1:0])
dst[191:160] := SELECT4(a[255:128], imm8[3:2])
dst[223:192] := SELECT4(b[255:128], imm8[5:4])
dst[255:224] := SELECT4(b[255:128], imm8[7:6])
dst[287:256] := SELECT4(a[383:256], imm8[1:0])
dst[319:288] := SELECT4(a[383:256], imm8[3:2])
dst[351:320] := SELECT4(b[383:256], imm8[5:4])
dst[383:352] := SELECT4(b[383:256], imm8[7:6])
dst[415:384] := SELECT4(a[511:384], imm8[1:0])
dst[447:416] := SELECT4(a[511:384], imm8[3:2])
dst[479:448] := SELECT4(b[511:384], imm8[5:4])
dst[511:480] := SELECT4(b[511:384], imm8[7:6])
dst[MAX:512] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm512_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512SinPd 

func M512SinPd(a x86.M512d) (dst x86.M512d)

M512SinPd: Compute the sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SIN(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sin_pd'. Requires AVX512F.

func M512SinPs 

func M512SinPs(a x86.M512) (dst x86.M512)

M512SinPs: Compute the sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SIN(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sin_ps'. Requires AVX512F.

func M512SincosPd 

func M512SincosPd(cos_res *x86.M512d, a x86.M512d) (dst x86.M512d)

M512SincosPd: Computes the sine and cosine of the packed double-precision (64-bit) floating-point elements in 'a' and stores the results of the sine computation in 'dst' and the results of the cosine computation in 'cos_res'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SIN(a[i+63:i])
	cos_res[i+63:i] := COS(a[i+63:i])
ENDFOR
dst[MAX:512] := 0
cos_res[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sincos_pd'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func M512SincosPs 

func M512SincosPs(cos_res *x86.M512, a x86.M512) (dst x86.M512)

M512SincosPs: Computes the sine and cosine of the packed single-precision (32-bit) floating-point elements in 'a' and stores the results of the sine computation in 'dst' and the results of the cosine computation in 'cos_res'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SIN(a[i+31:i])
	cos_res[i+31:i] := COS(a[i+31:i])
ENDFOR
dst[MAX:512] := 0
cos_res[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sincos_ps'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func M512SindPd 

func M512SindPd(a x86.M512d) (dst x86.M512d)

M512SindPd: Compute the sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SIND(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sind_pd'. Requires AVX512F.

func M512SindPs 

func M512SindPs(a x86.M512) (dst x86.M512)

M512SindPs: Compute the sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SIND(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sind_ps'. Requires AVX512F.

func M512SinhPd 

func M512SinhPd(a x86.M512d) (dst x86.M512d)

M512SinhPd: Compute the hyperbolic sine of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SINH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sinh_pd'. Requires AVX512F.

func M512SinhPs 

func M512SinhPs(a x86.M512) (dst x86.M512)

M512SinhPs: Compute the hyperbolic sine of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SINH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_sinh_ps'. Requires AVX512F.

func M512SllEpi32 

func M512SllEpi32(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	IF count[63:0] > 31
		dst[i+31:i] := 0
	ELSE
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm512_sll_epi32'. Requires AVX512F.

func M512SllEpi64 

func M512SllEpi64(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF count[63:0] > 63
		dst[i+63:i] := 0
	ELSE
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_sll_epi64'. Requires AVX512F.

func M512SlliEpi64 

func M512SlliEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512SlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF imm8[7:0] > 63
		dst[i+63:i] := 0
	ELSE
		dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm512_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512SllvEpi64 

func M512SllvEpi64(a x86.M512i, count x86.M512i) (dst x86.M512i)

M512SllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm512_sllv_epi64'. Requires AVX512F.

func M512SqrtPd 

func M512SqrtPd(a x86.M512d) (dst x86.M512d)

M512SqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SQRT(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_sqrt_pd'. Requires AVX512F.

func M512SqrtPs 

func M512SqrtPs(a x86.M512) (dst x86.M512)

M512SqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := SQRT(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_sqrt_ps'. Requires AVX512F.

func M512SqrtRoundPd 

func M512SqrtRoundPd(a x86.M512d, rounding int) (dst x86.M512d)

M512SqrtRoundPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE.

		FOR j := 0 to 7
			i := j*64
			dst[i+63:i] := SQRT(a[i+63:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm512_sqrt_round_pd'. Requires AVX512F.

func M512SqrtRoundPs 

func M512SqrtRoundPs(a x86.M512, rounding int) (dst x86.M512)

M512SqrtRoundPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE.

		FOR j := 0 to 15
			i := j*32
			dst[i+31:i] := SQRT(a[i+31:i])
		ENDFOR
		dst[MAX:512] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm512_sqrt_round_ps'. Requires AVX512F.

func M512SraEpi32 

func M512SraEpi32(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	IF count[63:0] > 31
		dst[i+31:i] := SignBit
	ELSE
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm512_sra_epi32'. Requires AVX512F.

func M512SraEpi64 

func M512SraEpi64(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF count[63:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_sra_epi64'. Requires AVX512F.

func M512SraiEpi64 

func M512SraiEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512SraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF imm8[7:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm512_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512SravEpi64 

func M512SravEpi64(a x86.M512i, count x86.M512i) (dst x86.M512i)

M512SravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm512_srav_epi64'. Requires AVX512F.

func M512SrlEpi32 

func M512SrlEpi32(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	IF count[63:0] > 31
		dst[i+31:i] := 0
	ELSE
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm512_srl_epi32'. Requires AVX512F.

func M512SrlEpi64 

func M512SrlEpi64(a x86.M512i, count x86.M128i) (dst x86.M512i)

M512SrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF count[63:0] > 63
		dst[i+63:i] := 0
	ELSE
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_srl_epi64'. Requires AVX512F.

func M512SrliEpi64 

func M512SrliEpi64(a x86.M512i, imm8 byte) (dst x86.M512i)

M512SrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	IF imm8[7:0] > 63
		dst[i+63:i] := 0
	ELSE
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
	FI
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm512_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512SrlvEpi64 

func M512SrlvEpi64(a x86.M512i, count x86.M512i) (dst x86.M512i)

M512SrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm512_srlv_epi64'. Requires AVX512F.

func M512SubEpi64 

func M512SubEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512SubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := a[i+63:i] - b[i+63:i]
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm512_sub_epi64'. Requires AVX512F.

func M512SvmlRoundPd 

func M512SvmlRoundPd(a x86.M512d) (dst x86.M512d)

M512SvmlRoundPd: Round the packed double-precision (64-bit) floating-point elements in 'a' to the nearest integer value, and store the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := ROUND(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_svml_round_pd'. Requires AVX512F.

func M512TanPd 

func M512TanPd(a x86.M512d) (dst x86.M512d)

M512TanPd: Compute the tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := TAN(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tan_pd'. Requires AVX512F.

func M512TanPs 

func M512TanPs(a x86.M512) (dst x86.M512)

M512TanPs: Compute the tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := TAN(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tan_ps'. Requires AVX512F.

func M512TandPd 

func M512TandPd(a x86.M512d) (dst x86.M512d)

M512TandPd: Compute the tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := TAND(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tand_pd'. Requires AVX512F.

func M512TandPs 

func M512TandPs(a x86.M512) (dst x86.M512)

M512TandPs: Compute the tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in degrees, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := TAND(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tand_ps'. Requires AVX512F.

func M512TanhPd 

func M512TanhPd(a x86.M512d) (dst x86.M512d)

M512TanhPd: Compute the hyperbolic tangent of packed double-precision (64-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := TANH(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tanh_pd'. Requires AVX512F.

func M512TanhPs 

func M512TanhPs(a x86.M512) (dst x86.M512)

M512TanhPs: Compute the hyperbolic tangent of packed single-precision (32-bit) floating-point elements in 'a' expressed in radians, and store the results in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := TANH(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_tanh_ps'. Requires AVX512F.

func M512TernarylogicEpi32 

func M512TernarylogicEpi32(a x86.M512i, b x86.M512i, c x86.M512i, imm8 byte) (dst x86.M512i)

M512TernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 15
	i := j*32
	FOR h := 0 to 31
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm512_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512TernarylogicEpi64 

func M512TernarylogicEpi64(a x86.M512i, b x86.M512i, c x86.M512i, imm8 byte) (dst x86.M512i)

M512TernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 7
	i := j*64
	FOR h := 0 to 63
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:512] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm512_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func M512TestEpi64Mask 

func M512TestEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512TestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is non-zero. 

FOR j := 0 to 7
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm512_test_epi64_mask'. Requires AVX512F.

func M512TestnEpi32Mask 

func M512TestnEpi32Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask16)

M512TestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 15
	i := j*32
	k[j] := ((a[i+31:i] AND b[i+31:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:16] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm512_testn_epi32_mask'. Requires AVX512F.

func M512TestnEpi64Mask 

func M512TestnEpi64Mask(a x86.M512i, b x86.M512i) (dst x86.Mmask8)

M512TestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 7
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:8] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm512_testn_epi64_mask'. Requires AVX512F.

func M512TruncPd 

func M512TruncPd(a x86.M512d) (dst x86.M512d)

M512TruncPd: Truncate the packed double-precision (64-bit) floating-point elements in 'a', and store the results as packed double-precision floating-point elements in 'dst'. 

FOR j := 0 to 7
	i := j*64
	dst[i+63:i] := TRUNCATE(a[i+63:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_trunc_pd'. Requires AVX512F.

func M512TruncPs 

func M512TruncPs(a x86.M512) (dst x86.M512)

M512TruncPs: Truncate the packed single-precision (32-bit) floating-point elements in 'a', and store the results as packed single-precision floating-point elements in 'dst'. 

FOR j := 0 to 15
	i := j*32
	dst[i+31:i] := TRUNCATE(a[i+31:i])
ENDFOR
dst[MAX:512] := 0

Instruction: '...'. Intrinsic: '_mm512_trunc_ps'. Requires AVX512F.

func M512Undefined 

func M512Undefined() (dst x86.M512)

M512Undefined: Return vector of type __m512 with undefined elements.

Instruction: ”. Intrinsic: '_mm512_undefined'. Requires AVX512F.

func M512UndefinedEpi32 

func M512UndefinedEpi32() (dst x86.M512i)

M512UndefinedEpi32: Return vector of type __m512i with undefined elements.

Instruction: ”. Intrinsic: '_mm512_undefined_epi32'. Requires AVX512F.

func M512UndefinedPd 

func M512UndefinedPd() (dst x86.M512d)

M512UndefinedPd: Return vector of type __m512d with undefined elements.

Instruction: ”. Intrinsic: '_mm512_undefined_pd'. Requires AVX512F.

func M512UndefinedPs 

func M512UndefinedPs() (dst x86.M512)

M512UndefinedPs: Return vector of type __m512 with undefined elements.

Instruction: ”. Intrinsic: '_mm512_undefined_ps'. Requires AVX512F.

func M512UnpackhiEpi32 

func M512UnpackhiEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512UnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm512_unpackhi_epi32'. Requires AVX512F.

func M512UnpackhiEpi64 

func M512UnpackhiEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512UnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm512_unpackhi_epi64'. Requires AVX512F.

func M512UnpackhiPd 

func M512UnpackhiPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512UnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_HIGH_QWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_HIGH_QWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_HIGH_QWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm512_unpackhi_pd'. Requires AVX512F.

func M512UnpackhiPs 

func M512UnpackhiPs(a x86.M512, b x86.M512) (dst x86.M512)

M512UnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_HIGH_DWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_HIGH_DWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_HIGH_DWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm512_unpackhi_ps'. Requires AVX512F.

func M512UnpackloEpi32 

func M512UnpackloEpi32(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512UnpackloEpi32: Unpack and interleave 32-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm512_unpacklo_epi32'. Requires AVX512F.

func M512UnpackloEpi64 

func M512UnpackloEpi64(a x86.M512i, b x86.M512i) (dst x86.M512i)

M512UnpackloEpi64: Unpack and interleave 64-bit integers from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm512_unpacklo_epi64'. Requires AVX512F.

func M512UnpackloPd 

func M512UnpackloPd(a x86.M512d, b x86.M512d) (dst x86.M512d)

M512UnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_QWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_QWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_QWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm512_unpacklo_pd'. Requires AVX512F.

func M512UnpackloPs 

func M512UnpackloPs(a x86.M512, b x86.M512) (dst x86.M512)

M512UnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in 'a' and 'b', and store the results in 'dst'. 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])
dst[255:128] := INTERLEAVE_DWORDS(a[255:128], b[255:128])
dst[383:256] := INTERLEAVE_DWORDS(a[383:256], b[383:256])
dst[511:384] := INTERLEAVE_DWORDS(a[511:384], b[511:384])
dst[MAX:512] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm512_unpacklo_ps'. Requires AVX512F.

func Mask2Permutex2varEpi32 

func Mask2Permutex2varEpi32(a x86.M128i, idx x86.M128i, k x86.Mmask8, b x86.M128i) (dst x86.M128i)

Mask2Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+2] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2D'. Intrinsic: '_mm_mask2_permutex2var_epi32'. Requires AVX512F.

func Mask2Permutex2varEpi64 

func Mask2Permutex2varEpi64(a x86.M128i, idx x86.M128i, k x86.Mmask8, b x86.M128i) (dst x86.M128i)

Mask2Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2Q'. Intrinsic: '_mm_mask2_permutex2var_epi64'. Requires AVX512F.

func Mask2Permutex2varPd 

func Mask2Permutex2varPd(a x86.M128d, idx x86.M128i, k x86.Mmask8, b x86.M128d) (dst x86.M128d)

Mask2Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set) 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := idx[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PD'. Intrinsic: '_mm_mask2_permutex2var_pd'. Requires AVX512F.

func Mask2Permutex2varPs 

func Mask2Permutex2varPs(a x86.M128, idx x86.M128i, k x86.Mmask8, b x86.M128) (dst x86.M128)

Mask2Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'idx' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+2] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := idx[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PS'. Intrinsic: '_mm_mask2_permutex2var_ps'. Requires AVX512F.

func Mask3FmaddPd 

func Mask3FmaddPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm_mask3_fmadd_pd'. Requires AVX512F.

func Mask3FmaddPs 

func Mask3FmaddPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm_mask3_fmadd_ps'. Requires AVX512F.

func Mask3FmaddRoundSd 

func Mask3FmaddRoundSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8, rounding int) (dst x86.M128d)

Mask3FmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := c[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_mask3_fmadd_round_sd'. Requires AVX512F.

func Mask3FmaddRoundSs 

func Mask3FmaddRoundSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8, rounding int) (dst x86.M128)

Mask3FmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := c[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_mask3_fmadd_round_ss'. Requires AVX512F.

func Mask3FmaddSd 

func Mask3FmaddSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := c[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_mask3_fmadd_sd'. Requires AVX512F.

func Mask3FmaddSs 

func Mask3FmaddSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := c[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_mask3_fmadd_ss'. Requires AVX512F.

func Mask3FmaddsubPd 

func Mask3FmaddsubPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm_mask3_fmaddsub_pd'. Requires AVX512F.

func Mask3FmaddsubPs 

func Mask3FmaddsubPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm_mask3_fmaddsub_ps'. Requires AVX512F.

func Mask3FmsubPd 

func Mask3FmsubPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm_mask3_fmsub_pd'. Requires AVX512F.

func Mask3FmsubPs 

func Mask3FmsubPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm_mask3_fmsub_ps'. Requires AVX512F.

func Mask3FmsubRoundSd 

func Mask3FmsubRoundSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8, rounding int) (dst x86.M128d)

Mask3FmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := c[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_mask3_fmsub_round_sd'. Requires AVX512F.

func Mask3FmsubRoundSs 

func Mask3FmsubRoundSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8, rounding int) (dst x86.M128)

Mask3FmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := c[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_mask3_fmsub_round_ss'. Requires AVX512F.

func Mask3FmsubSd 

func Mask3FmsubSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := c[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_mask3_fmsub_sd'. Requires AVX512F.

func Mask3FmsubSs 

func Mask3FmsubSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := c[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_mask3_fmsub_ss'. Requires AVX512F.

func Mask3FmsubaddPd 

func Mask3FmsubaddPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm_mask3_fmsubadd_pd'. Requires AVX512F.

func Mask3FmsubaddPs 

func Mask3FmsubaddPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm_mask3_fmsubadd_ps'. Requires AVX512F.

func Mask3FnmaddPd 

func Mask3FnmaddPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm_mask3_fnmadd_pd'. Requires AVX512F.

func Mask3FnmaddPs 

func Mask3FnmaddPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm_mask3_fnmadd_ps'. Requires AVX512F.

func Mask3FnmaddRoundSd 

func Mask3FnmaddRoundSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8, rounding int) (dst x86.M128d)

Mask3FnmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := c[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SD, VFNMADD213SD, VFNMADD231SD'. Intrinsic: '_mm_mask3_fnmadd_round_sd'. Requires AVX512F.

func Mask3FnmaddRoundSs 

func Mask3FnmaddRoundSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8, rounding int) (dst x86.M128)

Mask3FnmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := c[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_mask3_fnmadd_round_ss'. Requires AVX512F.

func Mask3FnmaddSd 

func Mask3FnmaddSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FnmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := c[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMADD132SD, VFNMADD213SD, VFNMADD231SD'. Intrinsic: '_mm_mask3_fnmadd_sd'. Requires AVX512F.

func Mask3FnmaddSs 

func Mask3FnmaddSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FnmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := c[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_mask3_fnmadd_ss'. Requires AVX512F.

func Mask3FnmsubPd 

func Mask3FnmsubPd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := c[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm_mask3_fnmsub_pd'. Requires AVX512F.

func Mask3FnmsubPs 

func Mask3FnmsubPs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := c[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm_mask3_fnmsub_ps'. Requires AVX512F.

func Mask3FnmsubRoundSd 

func Mask3FnmsubRoundSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8, rounding int) (dst x86.M128d)

Mask3FnmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := c[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_mask3_fnmsub_round_sd'. Requires AVX512F.

func Mask3FnmsubRoundSs 

func Mask3FnmsubRoundSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8, rounding int) (dst x86.M128)

Mask3FnmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', subtract the lower element in 'c' from the negated intermediate result, store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := c[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_mask3_fnmsub_round_ss'. Requires AVX512F.

func Mask3FnmsubSd 

func Mask3FnmsubSd(a x86.M128d, b x86.M128d, c x86.M128d, k x86.Mmask8) (dst x86.M128d)

Mask3FnmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := c[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_mask3_fnmsub_sd'. Requires AVX512F.

func Mask3FnmsubSs 

func Mask3FnmsubSs(a x86.M128, b x86.M128, c x86.M128, k x86.Mmask8) (dst x86.M128)

Mask3FnmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst' using writemask 'k' (elements are copied from 'c' when the corresponding mask bit is not set). 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := c[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_mask3_fnmsub_ss'. Requires AVX512F.

func MaskAbsEpi32 

func MaskAbsEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm_mask_abs_epi32'. Requires AVX512F.

func MaskAbsEpi64 

func MaskAbsEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm_mask_abs_epi64'. Requires AVX512F.

func MaskAddEpi32 

func MaskAddEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAddEpi32: Add packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPADDD'. Intrinsic: '_mm_mask_add_epi32'. Requires AVX512F.

func MaskAddEpi64 

func MaskAddEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm_mask_add_epi64'. Requires AVX512F.

func MaskAddRoundSd 

func MaskAddRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskAddRoundSd: Add the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] + b[63:0]
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VADDSD'. Intrinsic: '_mm_mask_add_round_sd'. Requires AVX512F.

func MaskAddRoundSs 

func MaskAddRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskAddRoundSs: Add the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] + b[31:0]
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VADDSS'. Intrinsic: '_mm_mask_add_round_ss'. Requires AVX512F.

func MaskAddSd 

func MaskAddSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskAddSd: Add the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] + b[63:0]
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VADDSD'. Intrinsic: '_mm_mask_add_sd'. Requires AVX512F.

func MaskAddSs 

func MaskAddSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskAddSs: Add the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] + b[31:0]
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VADDSS'. Intrinsic: '_mm_mask_add_ss'. Requires AVX512F.

func MaskAndEpi32 

func MaskAndEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAndEpi32: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] AND b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDD'. Intrinsic: '_mm_mask_and_epi32'. Requires AVX512F.

func MaskAndEpi64 

func MaskAndEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAndEpi64: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] AND b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDQ'. Intrinsic: '_mm_mask_and_epi64'. Requires AVX512F.

func MaskAndnotEpi32 

func MaskAndnotEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAndnotEpi32: Compute the bitwise AND NOT of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ((NOT a[i+31:i]) AND b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDND'. Intrinsic: '_mm_mask_andnot_epi32'. Requires AVX512F.

func MaskAndnotEpi64 

func MaskAndnotEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskAndnotEpi64: Compute the bitwise AND NOT of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ((NOT a[i+63:i]) AND b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDNQ'. Intrinsic: '_mm_mask_andnot_epi64'. Requires AVX512F.

func MaskBlendEpi32 

func MaskBlendEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskBlendEpi32: Blend packed 32-bit integers from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := b[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBLENDMD'. Intrinsic: '_mm_mask_blend_epi32'. Requires AVX512F.

func MaskBlendEpi64 

func MaskBlendEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskBlendEpi64: Blend packed 64-bit integers from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := b[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBLENDMQ'. Intrinsic: '_mm_mask_blend_epi64'. Requires AVX512F.

func MaskBlendPd 

func MaskBlendPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskBlendPd: Blend packed double-precision (64-bit) floating-point elements from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := b[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VBLENDMPD'. Intrinsic: '_mm_mask_blend_pd'. Requires AVX512F.

func MaskBlendPs 

func MaskBlendPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskBlendPs: Blend packed single-precision (32-bit) floating-point elements from 'a' and 'b' using control mask 'k', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := b[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VBLENDMPS'. Intrinsic: '_mm_mask_blend_ps'. Requires AVX512F.

func MaskBroadcastdEpi32 

func MaskBroadcastdEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm_mask_broadcastd_epi32'. Requires AVX512F.

func MaskBroadcastqEpi64 

func MaskBroadcastqEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm_mask_broadcastq_epi64'. Requires AVX512F.

func MaskBroadcastssPs 

func MaskBroadcastssPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm_mask_broadcastss_ps'. Requires AVX512F.

func MaskCmpEpi32Mask 

func MaskCmpEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

MaskCmpEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmp_epi32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpEpi64Mask 

func MaskCmpEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

MaskCmpEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmp_epi64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpEpu32Mask 

func MaskCmpEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

MaskCmpEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmp_epu32_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpEpu64Mask 

func MaskCmpEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.Mmask8)

MaskCmpEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _MM_CMPINT_EQ
1: OP := _MM_CMPINT_LT
2: OP := _MM_CMPINT_LE
3: OP := _MM_CMPINT_FALSE
4: OP := _MM_CMPINT_NEQ
5: OP := _MM_CMPINT_NLT
6: OP := _MM_CMPINT_NLE
7: OP := _MM_CMPINT_TRUE
ESAC
FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmp_epu64_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpPdMask 

func MaskCmpPdMask(k1 x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.Mmask8)

MaskCmpPdMask: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k' using zeromask 'k1' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] OP b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VCMPPD'. Intrinsic: '_mm_mask_cmp_pd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpPsMask 

func MaskCmpPsMask(k1 x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.Mmask8)

MaskCmpPsMask: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the results in mask vector 'k' using zeromask 'k1' (elements are zeroed out when the corresponding mask bit is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC
FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] OP b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VCMPPS'. Intrinsic: '_mm_mask_cmp_ps_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpRoundSdMask 

func MaskCmpRoundSdMask(k1 x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte, sae int) (dst x86.Mmask8)

MaskCmpRoundSdMask: Compare the lower double-precision (64-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k' using zeromask 'k1' (the element is zeroed out when mask bit 0 is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	IF k1[0]
		k[0] := ( a[63:0] OP b[63:0] ) ? 1 : 0
	ELSE
		k[0] := 0
	FI
	k[MAX:1] := 0

Instruction: 'VCMPSD'. Intrinsic: '_mm_mask_cmp_round_sd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpRoundSsMask 

func MaskCmpRoundSsMask(k1 x86.Mmask8, a x86.M128, b x86.M128, imm8 byte, sae int) (dst x86.Mmask8)

MaskCmpRoundSsMask: Compare the lower single-precision (32-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k' using zeromask 'k1' (the element is zeroed out when mask bit 0 is not set). 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	CASE (imm8[7:0]) OF
	0: OP := _CMP_EQ_OQ
	1: OP := _CMP_LT_OS
	2: OP := _CMP_LE_OS
	3: OP := _CMP_UNORD_Q
	4: OP := _CMP_NEQ_UQ
	5: OP := _CMP_NLT_US
	6: OP := _CMP_NLE_US
	7: OP := _CMP_ORD_Q
	8: OP := _CMP_EQ_UQ
	9: OP := _CMP_NGE_US
	10: OP := _CMP_NGT_US
	11: OP := _CMP_FALSE_OQ
	12: OP := _CMP_NEQ_OQ
	13: OP := _CMP_GE_OS
	14: OP := _CMP_GT_OS
	15: OP := _CMP_TRUE_UQ
	16: OP := _CMP_EQ_OS
	17: OP := _CMP_LT_OQ
	18: OP := _CMP_LE_OQ
	19: OP := _CMP_UNORD_S
	20: OP := _CMP_NEQ_US
	21: OP := _CMP_NLT_UQ
	22: OP := _CMP_NLE_UQ
	23: OP := _CMP_ORD_S
	24: OP := _CMP_EQ_US
	25: OP := _CMP_NGE_UQ
	26: OP := _CMP_NGT_UQ
	27: OP := _CMP_FALSE_OS
	28: OP := _CMP_NEQ_OS
	29: OP := _CMP_GE_OQ
	30: OP := _CMP_GT_OQ
	31: OP := _CMP_TRUE_US
	ESAC

	IF k1[0]
		k[0] := ( a[31:0] OP b[31:0] ) ? 1 : 0
	ELSE
		k[0] := 0
	FI
	k[MAX:1] := 0

Instruction: 'VCMPSS'. Intrinsic: '_mm_mask_cmp_round_ss_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpSdMask 

func MaskCmpSdMask(k1 x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.Mmask8)

MaskCmpSdMask: Compare the lower double-precision (64-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k' using zeromask 'k1' (the element is zeroed out when mask bit 0 is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC

IF k1[0]
	k[0] := ( a[63:0] OP b[63:0] ) ? 1 : 0
ELSE
	k[0] := 0
FI
k[MAX:1] := 0

Instruction: 'VCMPSD'. Intrinsic: '_mm_mask_cmp_sd_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpSsMask 

func MaskCmpSsMask(k1 x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.Mmask8)

MaskCmpSsMask: Compare the lower single-precision (32-bit) floating-point element in 'a' and 'b' based on the comparison operand specified by 'imm8', and store the result in mask vector 'k' using zeromask 'k1' (the element is zeroed out when mask bit 0 is not set). 

CASE (imm8[7:0]) OF
0: OP := _CMP_EQ_OQ
1: OP := _CMP_LT_OS
2: OP := _CMP_LE_OS
3: OP := _CMP_UNORD_Q
4: OP := _CMP_NEQ_UQ
5: OP := _CMP_NLT_US
6: OP := _CMP_NLE_US
7: OP := _CMP_ORD_Q
8: OP := _CMP_EQ_UQ
9: OP := _CMP_NGE_US
10: OP := _CMP_NGT_US
11: OP := _CMP_FALSE_OQ
12: OP := _CMP_NEQ_OQ
13: OP := _CMP_GE_OS
14: OP := _CMP_GT_OS
15: OP := _CMP_TRUE_UQ
16: OP := _CMP_EQ_OS
17: OP := _CMP_LT_OQ
18: OP := _CMP_LE_OQ
19: OP := _CMP_UNORD_S
20: OP := _CMP_NEQ_US
21: OP := _CMP_NLT_UQ
22: OP := _CMP_NLE_UQ
23: OP := _CMP_ORD_S
24: OP := _CMP_EQ_US
25: OP := _CMP_NGE_UQ
26: OP := _CMP_NGT_UQ
27: OP := _CMP_FALSE_OS
28: OP := _CMP_NEQ_OS
29: OP := _CMP_GE_OQ
30: OP := _CMP_GT_OQ
31: OP := _CMP_TRUE_US
ESAC

IF k1[0]
	k[0] := ( a[31:0] OP b[31:0] ) ? 1 : 0
ELSE
	k[0] := 0
FI
k[MAX:1] := 0

Instruction: 'VCMPSS'. Intrinsic: '_mm_mask_cmp_ss_mask'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskCmpeqEpi32Mask 

func MaskCmpeqEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpeqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmpeq_epi32_mask'. Requires AVX512F.

func MaskCmpeqEpi64Mask 

func MaskCmpeqEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpeqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmpeq_epi64_mask'. Requires AVX512F.

func MaskCmpeqEpu32Mask 

func MaskCmpeqEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpeqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] == b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmpeq_epu32_mask'. Requires AVX512F.

func MaskCmpeqEpu64Mask 

func MaskCmpeqEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpeqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for equality, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] == b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmpeq_epu64_mask'. Requires AVX512F.

func MaskCmpgeEpi32Mask 

func MaskCmpgeEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgeEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmpge_epi32_mask'. Requires AVX512F.

func MaskCmpgeEpi64Mask 

func MaskCmpgeEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgeEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmpge_epi64_mask'. Requires AVX512F.

func MaskCmpgeEpu32Mask 

func MaskCmpgeEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgeEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] >= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmpge_epu32_mask'. Requires AVX512F.

func MaskCmpgeEpu64Mask 

func MaskCmpgeEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgeEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] >= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmpge_epu64_mask'. Requires AVX512F.

func MaskCmpgtEpi32Mask 

func MaskCmpgtEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgtEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmpgt_epi32_mask'. Requires AVX512F.

func MaskCmpgtEpi64Mask 

func MaskCmpgtEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgtEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmpgt_epi64_mask'. Requires AVX512F.

func MaskCmpgtEpu32Mask 

func MaskCmpgtEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgtEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] > b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmpgt_epu32_mask'. Requires AVX512F.

func MaskCmpgtEpu64Mask 

func MaskCmpgtEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpgtEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for greater-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] > b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmpgt_epu64_mask'. Requires AVX512F.

func MaskCmpleEpi32Mask 

func MaskCmpleEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpleEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmple_epi32_mask'. Requires AVX512F.

func MaskCmpleEpi64Mask 

func MaskCmpleEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpleEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmple_epi64_mask'. Requires AVX512F.

func MaskCmpleEpu32Mask 

func MaskCmpleEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpleEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] <= b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmple_epu32_mask'. Requires AVX512F.

func MaskCmpleEpu64Mask 

func MaskCmpleEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpleEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than-or-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] <= b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmple_epu64_mask'. Requires AVX512F.

func MaskCmpltEpi32Mask 

func MaskCmpltEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpltEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmplt_epi32_mask'. Requires AVX512F.

func MaskCmpltEpi64Mask 

func MaskCmpltEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpltEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmplt_epi64_mask'. Requires AVX512F.

func MaskCmpltEpu32Mask 

func MaskCmpltEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpltEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] < b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmplt_epu32_mask'. Requires AVX512F.

func MaskCmpltEpu64Mask 

func MaskCmpltEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpltEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for less-than, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] < b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmplt_epu64_mask'. Requires AVX512F.

func MaskCmpneqEpi32Mask 

func MaskCmpneqEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpneqEpi32Mask: Compare packed 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPD'. Intrinsic: '_mm_mask_cmpneq_epi32_mask'. Requires AVX512F.

func MaskCmpneqEpi64Mask 

func MaskCmpneqEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpneqEpi64Mask: Compare packed 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPQ'. Intrinsic: '_mm_mask_cmpneq_epi64_mask'. Requires AVX512F.

func MaskCmpneqEpu32Mask 

func MaskCmpneqEpu32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpneqEpu32Mask: Compare packed unsigned 32-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ( a[i+31:i] != b[i+31:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPCMPUD'. Intrinsic: '_mm_mask_cmpneq_epu32_mask'. Requires AVX512F.

func MaskCmpneqEpu64Mask 

func MaskCmpneqEpu64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskCmpneqEpu64Mask: Compare packed unsigned 64-bit integers in 'a' and 'b' for not-equal, and store the results in mask vector 'k1' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ( a[i+63:i] != b[i+63:i] ) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPCMPUQ'. Intrinsic: '_mm_mask_cmpneq_epu64_mask'. Requires AVX512F.

func MaskCompressEpi32 

func MaskCompressEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := src[127:m]
dst[MAX:128] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm_mask_compress_epi32'. Requires AVX512F.

func MaskCompressEpi64 

func MaskCompressEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := src[127:m]
dst[MAX:128] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm_mask_compress_epi64'. Requires AVX512F.

func MaskCompressPd 

func MaskCompressPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 64
m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := src[127:m]
dst[MAX:128] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm_mask_compress_pd'. Requires AVX512F.

func MaskCompressPs 

func MaskCompressPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in writemask 'k') to 'dst', and pass through the remaining elements from 'src'. 

size := 32
m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := src[127:m]
dst[MAX:128] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm_mask_compress_ps'. Requires AVX512F.

func MaskCvtRoundpsPh 

func MaskCvtRoundpsPh(src x86.M128i, k x86.Mmask8, a x86.M128, rounding int) (dst x86.M128i)

MaskCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 3
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:64] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm_mask_cvt_roundps_ph'. Requires AVX512F.

func MaskCvtRoundsdSs 

func MaskCvtRoundsdSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128d, rounding int) (dst x86.M128)

MaskCvtRoundsdSs: Convert the lower double-precision (64-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := Convert_FP64_To_FP32(b[63:0])
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:31]
		dst[MAX:64] := 0

Instruction: 'VCVTSD2SS'. Intrinsic: '_mm_mask_cvt_roundsd_ss'. Requires AVX512F.

func MaskCvtRoundssSd 

func MaskCvtRoundssSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128, rounding int) (dst x86.M128d)

MaskCvtRoundssSd: Convert the lower single-precision (32-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := Convert_FP32_To_FP64(b[31:0])
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:64] := 0

Instruction: 'VCVTSS2SD'. Intrinsic: '_mm_mask_cvt_roundss_sd'. Requires AVX512F.

func MaskCvtepi16Epi32 

func MaskCvtepi16Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	l := j*16
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm_mask_cvtepi16_epi32'. Requires AVX512F.

func MaskCvtepi16Epi64 

func MaskCvtepi16Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi16Epi64: Sign extend packed 16-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm_mask_cvtepi16_epi64'. Requires AVX512F.

func MaskCvtepi32Epi16 

func MaskCvtepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm_mask_cvtepi32_epi16'. Requires AVX512F.

func MaskCvtepi32Epi64 

func MaskCvtepi32Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm_mask_cvtepi32_epi64'. Requires AVX512F.

func MaskCvtepi32Epi8 

func MaskCvtepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm_mask_cvtepi32_epi8'. Requires AVX512F.

func MaskCvtepi32Pd 

func MaskCvtepi32Pd(src x86.M128d, k x86.Mmask8, a x86.M128i) (dst x86.M128d)

MaskCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := src[m+63:m]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm_mask_cvtepi32_pd'. Requires AVX512F.

func MaskCvtepi32Ps 

func MaskCvtepi32Ps(src x86.M128, k x86.Mmask8, a x86.M128i) (dst x86.M128)

MaskCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm_mask_cvtepi32_ps'. Requires AVX512F.

func MaskCvtepi64Epi16 

func MaskCvtepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm_mask_cvtepi64_epi16'. Requires AVX512F.

func MaskCvtepi64Epi32 

func MaskCvtepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm_mask_cvtepi64_epi32'. Requires AVX512F.

func MaskCvtepi64Epi8 

func MaskCvtepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm_mask_cvtepi64_epi8'. Requires AVX512F.

func MaskCvtepi8Epi32 

func MaskCvtepi8Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi8Epi32: Sign extend packed 8-bit integers in the low 4 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm_mask_cvtepi8_epi32'. Requires AVX512F.

func MaskCvtepi8Epi64 

func MaskCvtepi8Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepi8Epi64: Sign extend packed 8-bit integers in the low 2 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm_mask_cvtepi8_epi64'. Requires AVX512F.

func MaskCvtepu16Epi32 

func MaskCvtepu16Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm_mask_cvtepu16_epi32'. Requires AVX512F.

func MaskCvtepu16Epi64 

func MaskCvtepu16Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm_mask_cvtepu16_epi64'. Requires AVX512F.

func MaskCvtepu32Epi64 

func MaskCvtepu32Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm_mask_cvtepu32_epi64'. Requires AVX512F.

func MaskCvtepu32Pd 

func MaskCvtepu32Pd(src x86.M128d, k x86.Mmask8, a x86.M128i) (dst x86.M128d)

MaskCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm_mask_cvtepu32_pd'. Requires AVX512F.

func MaskCvtepu8Epi32 

func MaskCvtepu8Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in the low 4 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm_mask_cvtepu8_epi32'. Requires AVX512F.

func MaskCvtepu8Epi64 

func MaskCvtepu8Epi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 2 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm_mask_cvtepu8_epi64'. Requires AVX512F.

func MaskCvtpdEpi32 

func MaskCvtpdEpi32(src x86.M128i, k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm_mask_cvtpd_epi32'. Requires AVX512F.

func MaskCvtpdEpu32 

func MaskCvtpdEpu32(src x86.M128i, k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm_mask_cvtpd_epu32'. Requires AVX512F.

func MaskCvtpdPs 

func MaskCvtpdPs(src x86.M128, k x86.Mmask8, a x86.M128d) (dst x86.M128)

MaskCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm_mask_cvtpd_ps'. Requires AVX512F.

func MaskCvtphPs 

func MaskCvtphPs(src x86.M128, k x86.Mmask8, a x86.M128i) (dst x86.M128)

MaskCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm_mask_cvtph_ps'. Requires AVX512F.

func MaskCvtpsEpi32 

func MaskCvtpsEpi32(src x86.M128i, k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm_mask_cvtps_epi32'. Requires AVX512F.

func MaskCvtpsEpu32 

func MaskCvtpsEpu32(src x86.M128i, k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm_mask_cvtps_epu32'. Requires AVX512F.

func MaskCvtpsPh 

func MaskCvtpsPh(src x86.M128i, k x86.Mmask8, a x86.M128, rounding int) (dst x86.M128i)

MaskCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 3
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := src[i+15:i]
			FI
		ENDFOR
		dst[MAX:64] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm_mask_cvtps_ph'. Requires AVX512F.

func MaskCvtsdSs 

func MaskCvtsdSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128d) (dst x86.M128)

MaskCvtsdSs: Convert the lower double-precision (64-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[31:0] := Convert_FP64_To_FP32(b[63:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:31]
dst[MAX:64] := 0

Instruction: 'VCVTSD2SS'. Intrinsic: '_mm_mask_cvtsd_ss'. Requires AVX512F.

func MaskCvtsepi32Epi16 

func MaskCvtsepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm_mask_cvtsepi32_epi16'. Requires AVX512F.

func MaskCvtsepi32Epi8 

func MaskCvtsepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm_mask_cvtsepi32_epi8'. Requires AVX512F.

func MaskCvtsepi64Epi16 

func MaskCvtsepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm_mask_cvtsepi64_epi16'. Requires AVX512F.

func MaskCvtsepi64Epi32 

func MaskCvtsepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm_mask_cvtsepi64_epi32'. Requires AVX512F.

func MaskCvtsepi64Epi8 

func MaskCvtsepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm_mask_cvtsepi64_epi8'. Requires AVX512F.

func MaskCvtssSd 

func MaskCvtssSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128) (dst x86.M128d)

MaskCvtssSd: Convert the lower single-precision (32-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := Convert_FP32_To_FP64(b[31:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:64] := 0

Instruction: 'VCVTSS2SD'. Intrinsic: '_mm_mask_cvtss_sd'. Requires AVX512F.

func MaskCvttpdEpi32 

func MaskCvttpdEpi32(src x86.M128i, k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm_mask_cvttpd_epi32'. Requires AVX512F.

func MaskCvttpdEpu32 

func MaskCvttpdEpu32(src x86.M128i, k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm_mask_cvttpd_epu32'. Requires AVX512F.

func MaskCvttpsEpi32 

func MaskCvttpsEpi32(src x86.M128i, k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm_mask_cvttps_epi32'. Requires AVX512F.

func MaskCvttpsEpu32 

func MaskCvttpsEpu32(src x86.M128i, k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm_mask_cvttps_epu32'. Requires AVX512F.

func MaskCvtusepi32Epi16 

func MaskCvtusepi32Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm_mask_cvtusepi32_epi16'. Requires AVX512F.

func MaskCvtusepi32Epi8 

func MaskCvtusepi32Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm_mask_cvtusepi32_epi8'. Requires AVX512F.

func MaskCvtusepi64Epi16 

func MaskCvtusepi64Epi16(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := src[l+15:l]
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm_mask_cvtusepi64_epi16'. Requires AVX512F.

func MaskCvtusepi64Epi32 

func MaskCvtusepi64Epi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := src[l+31:l]
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm_mask_cvtusepi64_epi32'. Requires AVX512F.

func MaskCvtusepi64Epi8 

func MaskCvtusepi64Epi8(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := src[l+7:l]
	FI
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm_mask_cvtusepi64_epi8'. Requires AVX512F.

func MaskDivPd 

func MaskDivPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm_mask_div_pd'. Requires AVX512F.

func MaskDivPs 

func MaskDivPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm_mask_div_ps'. Requires AVX512F.

func MaskDivRoundSd 

func MaskDivRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskDivRoundSd: Divide the lower double-precision (64-bit) floating-point element in 'a' by the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] / b[63:0]
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VDIVSD'. Intrinsic: '_mm_mask_div_round_sd'. Requires AVX512F.

func MaskDivRoundSs 

func MaskDivRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskDivRoundSs: Divide the lower single-precision (32-bit) floating-point element in 'a' by the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] / b[31:0]
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VDIVSS'. Intrinsic: '_mm_mask_div_round_ss'. Requires AVX512F.

func MaskDivSd 

func MaskDivSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskDivSd: Divide the lower double-precision (64-bit) floating-point element in 'a' by the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] / b[63:0]
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VDIVSD'. Intrinsic: '_mm_mask_div_sd'. Requires AVX512F.

func MaskDivSs 

func MaskDivSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskDivSs: Divide the lower single-precision (32-bit) floating-point element in 'a' by the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] / b[31:0]
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VDIVSS'. Intrinsic: '_mm_mask_div_ss'. Requires AVX512F.

func MaskExpandEpi32 

func MaskExpandEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm_mask_expand_epi32'. Requires AVX512F.

func MaskExpandEpi64 

func MaskExpandEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm_mask_expand_epi64'. Requires AVX512F.

func MaskExpandPd 

func MaskExpandPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm_mask_expand_pd'. Requires AVX512F.

func MaskExpandPs 

func MaskExpandPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm_mask_expand_ps'. Requires AVX512F.

func MaskFixupimmPd 

func MaskFixupimmPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

MaskFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm_mask_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFixupimmPs 

func MaskFixupimmPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

MaskFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm_mask_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFixupimmRoundSd 

func MaskFixupimmRoundSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128i, imm8 byte, rounding int) (dst x86.M128d)

MaskFixupimmRoundSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		IF k[0]
			dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
		ELSE
			dst[63:0] := a[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_mask_fixupimm_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFixupimmRoundSs 

func MaskFixupimmRoundSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128i, imm8 byte, rounding int) (dst x86.M128)

MaskFixupimmRoundSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		IF k[0]
			dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
		ELSE
			dst[31:0] := a[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_mask_fixupimm_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFixupimmSd 

func MaskFixupimmSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

MaskFixupimmSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

IF k[0]
	dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
ELSE
	dst[63:0] := a[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_mask_fixupimm_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFixupimmSs 

func MaskFixupimmSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

MaskFixupimmSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

IF k[0]
	dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
ELSE
	dst[31:0] := a[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_mask_fixupimm_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskFmaddPd 

func MaskFmaddPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm_mask_fmadd_pd'. Requires AVX512F.

func MaskFmaddPs 

func MaskFmaddPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm_mask_fmadd_ps'. Requires AVX512F.

func MaskFmaddRoundSd 

func MaskFmaddRoundSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskFmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := a[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_mask_fmadd_round_sd'. Requires AVX512F.

func MaskFmaddRoundSs 

func MaskFmaddRoundSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskFmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := a[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_mask_fmadd_round_ss'. Requires AVX512F.

func MaskFmaddSd 

func MaskFmaddSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := a[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_mask_fmadd_sd'. Requires AVX512F.

func MaskFmaddSs 

func MaskFmaddSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := a[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_mask_fmadd_ss'. Requires AVX512F.

func MaskFmaddsubPd 

func MaskFmaddsubPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm_mask_fmaddsub_pd'. Requires AVX512F.

func MaskFmaddsubPs 

func MaskFmaddsubPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm_mask_fmaddsub_ps'. Requires AVX512F.

func MaskFmsubPd 

func MaskFmsubPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm_mask_fmsub_pd'. Requires AVX512F.

func MaskFmsubPs 

func MaskFmsubPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm_mask_fmsub_ps'. Requires AVX512F.

func MaskFmsubRoundSd 

func MaskFmsubRoundSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskFmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := a[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_mask_fmsub_round_sd'. Requires AVX512F.

func MaskFmsubRoundSs 

func MaskFmsubRoundSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskFmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := a[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_mask_fmsub_round_ss'. Requires AVX512F.

func MaskFmsubSd 

func MaskFmsubSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := a[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_mask_fmsub_sd'. Requires AVX512F.

func MaskFmsubSs 

func MaskFmsubSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := a[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_mask_fmsub_ss'. Requires AVX512F.

func MaskFmsubaddPd 

func MaskFmsubaddPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm_mask_fmsubadd_pd'. Requires AVX512F.

func MaskFmsubaddPs 

func MaskFmsubaddPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm_mask_fmsubadd_ps'. Requires AVX512F.

func MaskFnmaddPd 

func MaskFnmaddPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm_mask_fnmadd_pd'. Requires AVX512F.

func MaskFnmaddPs 

func MaskFnmaddPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm_mask_fnmadd_ps'. Requires AVX512F.

func MaskFnmaddRoundSd 

func MaskFnmaddRoundSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskFnmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := a[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SD, VFNMADD213SD, VFNMADD231SD'. Intrinsic: '_mm_mask_fnmadd_round_sd'. Requires AVX512F.

func MaskFnmaddRoundSs 

func MaskFnmaddRoundSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskFnmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := a[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_mask_fnmadd_round_ss'. Requires AVX512F.

func MaskFnmaddSd 

func MaskFnmaddSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFnmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := a[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMADD132SD, VFNMADD213SD, VFNMADD231SD'. Intrinsic: '_mm_mask_fnmadd_sd'. Requires AVX512F.

func MaskFnmaddSs 

func MaskFnmaddSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFnmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'a' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := a[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_mask_fnmadd_ss'. Requires AVX512F.

func MaskFnmsubPd 

func MaskFnmsubPd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm_mask_fnmsub_pd'. Requires AVX512F.

func MaskFnmsubPs 

func MaskFnmsubPs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm_mask_fnmsub_ps'. Requires AVX512F.

func MaskFnmsubRoundSd 

func MaskFnmsubRoundSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskFnmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := a[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_mask_fnmsub_round_sd'. Requires AVX512F.

func MaskFnmsubRoundSs 

func MaskFnmsubRoundSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskFnmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := a[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_mask_fnmsub_round_ss'. Requires AVX512F.

func MaskFnmsubSd 

func MaskFnmsubSd(a x86.M128d, k x86.Mmask8, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskFnmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := a[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_mask_fnmsub_sd'. Requires AVX512F.

func MaskFnmsubSs 

func MaskFnmsubSs(a x86.M128, k x86.Mmask8, b x86.M128, c x86.M128) (dst x86.M128)

MaskFnmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'c' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := a[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_mask_fnmsub_ss'. Requires AVX512F.

func MaskGetexpPd 

func MaskGetexpPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskGetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm_mask_getexp_pd'. Requires AVX512F.

func MaskGetexpPs 

func MaskGetexpPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskGetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm_mask_getexp_ps'. Requires AVX512F.

func MaskGetexpRoundSd 

func MaskGetexpRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskGetexpRoundSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := ConvertExpFP64(b[63:0])
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_mask_getexp_round_sd'. Requires AVX512F.

func MaskGetexpRoundSs 

func MaskGetexpRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskGetexpRoundSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := ConvertExpFP32(b[31:0])
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_mask_getexp_round_ss'. Requires AVX512F.

func MaskGetexpSd 

func MaskGetexpSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskGetexpSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

IF k[0]
	dst[63:0] := ConvertExpFP64(b[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_mask_getexp_sd'. Requires AVX512F.

func MaskGetexpSs 

func MaskGetexpSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskGetexpSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

IF k[0]
	dst[31:0] := ConvertExpFP32(b[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_mask_getexp_ss'. Requires AVX512F.

func MaskGetmantPd 

func MaskGetmantPd(src x86.M128d, k x86.Mmask8, a x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

MaskGetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 1
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := src[i+63:i]
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm_mask_getmant_pd'. Requires AVX512F.

func MaskGetmantPs 

func MaskGetmantPs(src x86.M128, k x86.Mmask8, a x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

MaskGetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := src[i+31:i]
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm_mask_getmant_ps'. Requires AVX512F.

func MaskGetmantRoundSd 

func MaskGetmantRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128d)

MaskGetmantRoundSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_mask_getmant_round_sd'. Requires AVX512F.

func MaskGetmantRoundSs 

func MaskGetmantRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128)

MaskGetmantRoundSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_mask_getmant_round_ss'. Requires AVX512F.

func MaskGetmantSd 

func MaskGetmantSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

MaskGetmantSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		IF k[0]
			dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_mask_getmant_sd'. Requires AVX512F.

func MaskGetmantSs 

func MaskGetmantSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

MaskGetmantSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		IF k[0]
			dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_mask_getmant_ss'. Requires AVX512F.

func MaskLoadSd 

func MaskLoadSd(src x86.M128d, k x86.Mmask8, mem_addr *float64) (dst x86.M128d)

MaskLoadSd: Load a double-precision (64-bit) floating-point element from memory into the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and set the upper element of 'dst' to zero. 'mem_addr' must be aligned on a 16-byte boundary or a general-protection exception may be generated. 

IF k[0]
	dst[63:0] := MEM[mem_addr+63:mem_addr]
ELSE
	dst[63:0] := src[63:0]
FI
dst[MAX:64] := 0

Instruction: 'VMOVSD'. Intrinsic: '_mm_mask_load_sd'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskLoadSs 

func MaskLoadSs(src x86.M128, k x86.Mmask8, mem_addr *float32) (dst x86.M128)

MaskLoadSs: Load a single-precision (32-bit) floating-point element from memory into the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and set the upper elements of 'dst' to zero. 'mem_addr' must be aligned on a 16-byte boundary or a general-protection exception may be generated. 

IF k[0]
	dst[31:0] := MEM[mem_addr+31:mem_addr]
ELSE
	dst[31:0] := src[31:0]
FI
dst[MAX:32] := 0

Instruction: 'VMOVSS'. Intrinsic: '_mm_mask_load_ss'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskMaxEpi32 

func MaskMaxEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMaxEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXSD'. Intrinsic: '_mm_mask_max_epi32'. Requires AVX512F.

func MaskMaxEpi64 

func MaskMaxEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm_mask_max_epi64'. Requires AVX512F.

func MaskMaxEpu32 

func MaskMaxEpu32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMaxEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXUD'. Intrinsic: '_mm_mask_max_epu32'. Requires AVX512F.

func MaskMaxEpu64 

func MaskMaxEpu64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm_mask_max_epu64'. Requires AVX512F.

func MaskMaxPd 

func MaskMaxPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm_mask_max_pd'. Requires AVX512F.

func MaskMaxPs 

func MaskMaxPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm_mask_max_ps'. Requires AVX512F.

func MaskMaxRoundSd 

func MaskMaxRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MaskMaxRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[63:0] := MAX(a[63:0], b[63:0])
	ELSE
		dst[63:0] := src[63:0]
	FI
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMAXSD'. Intrinsic: '_mm_mask_max_round_sd'. Requires AVX512F.

func MaskMaxRoundSs 

func MaskMaxRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, sae int) (dst x86.M128)

MaskMaxRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[31:0] := MAX(a[31:0], b[31:0])
	ELSE
		dst[31:0] := src[31:0]
	FI
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMAXSS'. Intrinsic: '_mm_mask_max_round_ss'. Requires AVX512F.

func MaskMaxSd 

func MaskMaxSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMaxSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := MAX(a[63:0], b[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMAXSD'. Intrinsic: '_mm_mask_max_sd'. Requires AVX512F.

func MaskMaxSs 

func MaskMaxSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMaxSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[31:0] := MAX(a[31:0], b[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMAXSS'. Intrinsic: '_mm_mask_max_ss'. Requires AVX512F.

func MaskMinEpi32 

func MaskMinEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMinEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINSD'. Intrinsic: '_mm_mask_min_epi32'. Requires AVX512F.

func MaskMinEpi64 

func MaskMinEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm_mask_min_epi64'. Requires AVX512F.

func MaskMinEpu32 

func MaskMinEpu32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMinEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINUD'. Intrinsic: '_mm_mask_min_epu32'. Requires AVX512F.

func MaskMinEpu64 

func MaskMinEpu64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm_mask_min_epu64'. Requires AVX512F.

func MaskMinPd 

func MaskMinPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm_mask_min_pd'. Requires AVX512F.

func MaskMinPs 

func MaskMinPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm_mask_min_ps'. Requires AVX512F.

func MaskMinRoundSd 

func MaskMinRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MaskMinRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[63:0] := MIN(a[63:0], b[63:0])
	ELSE
		dst[63:0] := src[63:0]
	FI
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMINSD'. Intrinsic: '_mm_mask_min_round_sd'. Requires AVX512F.

func MaskMinRoundSs 

func MaskMinRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, sae int) (dst x86.M128)

MaskMinRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[31:0] := MIN(a[31:0], b[31:0])
	ELSE
		dst[31:0] := src[31:0]
	FI
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMINSS'. Intrinsic: '_mm_mask_min_round_ss'. Requires AVX512F.

func MaskMinSd 

func MaskMinSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMinSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := MIN(a[63:0], b[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMINSD'. Intrinsic: '_mm_mask_min_sd'. Requires AVX512F.

func MaskMinSs 

func MaskMinSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMinSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[31:0] := MIN(a[31:0], b[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMINSS'. Intrinsic: '_mm_mask_min_ss'. Requires AVX512F.

func MaskMovEpi32 

func MaskMovEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskMovEpi32: Move packed 32-bit integers from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDQA32'. Intrinsic: '_mm_mask_mov_epi32'. Requires AVX512F.

func MaskMovEpi64 

func MaskMovEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskMovEpi64: Move packed 64-bit integers from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDQA64'. Intrinsic: '_mm_mask_mov_epi64'. Requires AVX512F.

func MaskMovPd 

func MaskMovPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskMovPd: Move packed double-precision (64-bit) floating-point elements from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVAPD'. Intrinsic: '_mm_mask_mov_pd'. Requires AVX512F.

func MaskMovPs 

func MaskMovPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskMovPs: Move packed single-precision (32-bit) floating-point elements from 'a' to 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVAPS'. Intrinsic: '_mm_mask_mov_ps'. Requires AVX512F.

func MaskMoveSd 

func MaskMoveSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMoveSd: Move the lower double-precision (64-bit) floating-point element from 'b' to the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := b[63:0]
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMOVSD'. Intrinsic: '_mm_mask_move_sd'. Requires AVX512F.

func MaskMoveSs 

func MaskMoveSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMoveSs: Move the lower single-precision (32-bit) floating-point element from 'b' to the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := b[31:0]
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMOVSS'. Intrinsic: '_mm_mask_move_ss'. Requires AVX512F.

func MaskMovedupPd 

func MaskMovedupPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm_mask_movedup_pd'. Requires AVX512F.

func MaskMovehdupPs 

func MaskMovehdupPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm_mask_movehdup_ps'. Requires AVX512F.

func MaskMoveldupPs 

func MaskMoveldupPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm_mask_moveldup_ps'. Requires AVX512F.

func MaskMulEpi32 

func MaskMulEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm_mask_mul_epi32'. Requires AVX512F.

func MaskMulEpu32 

func MaskMulEpu32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm_mask_mul_epu32'. Requires AVX512F.

func MaskMulPd 

func MaskMulPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMulPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm_mask_mul_pd'. Requires AVX512F.

func MaskMulPs 

func MaskMulPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMulPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). RM. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm_mask_mul_ps'. Requires AVX512F.

func MaskMulRoundSd 

func MaskMulRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskMulRoundSd: Multiply the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] * b[63:0]
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VMULSD'. Intrinsic: '_mm_mask_mul_round_sd'. Requires AVX512F.

func MaskMulRoundSs 

func MaskMulRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskMulRoundSs: Multiply the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] * b[31:0]
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VMULSS'. Intrinsic: '_mm_mask_mul_round_ss'. Requires AVX512F.

func MaskMulSd 

func MaskMulSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskMulSd: Multiply the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] * b[63:0]
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMULSD'. Intrinsic: '_mm_mask_mul_sd'. Requires AVX512F.

func MaskMulSs 

func MaskMulSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskMulSs: Multiply the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] * b[31:0]
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMULSS'. Intrinsic: '_mm_mask_mul_ss'. Requires AVX512F.

func MaskMulloEpi32 

func MaskMulloEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskMulloEpi32: Multiply the packed 32-bit integers in 'a' and 'b', producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		tmp[63:0] := a[i+31:i] * b[i+31:i]
		dst[i+31:i] := tmp[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULLD'. Intrinsic: '_mm_mask_mullo_epi32'. Requires AVX512F.

func MaskOrEpi32 

func MaskOrEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskOrEpi32: Compute the bitwise OR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] OR b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPORD'. Intrinsic: '_mm_mask_or_epi32'. Requires AVX512F.

func MaskOrEpi64 

func MaskOrEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskOrEpi64: Compute the bitwise OR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] OR b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPORQ'. Intrinsic: '_mm_mask_or_epi64'. Requires AVX512F.

func MaskPermutePd 

func MaskPermutePd(src x86.M128d, k x86.Mmask8, a x86.M128d, imm8 byte) (dst x86.M128d)

MaskPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm_mask_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskPermutePs 

func MaskPermutePs(src x86.M128, k x86.Mmask8, a x86.M128, imm8 byte) (dst x86.M128)

MaskPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm_mask_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskPermutevarPd 

func MaskPermutevarPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128i) (dst x86.M128d)

MaskPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm_mask_permutevar_pd'. Requires AVX512F.

func MaskPermutevarPs 

func MaskPermutevarPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128i) (dst x86.M128)

MaskPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm_mask_permutevar_ps'. Requires AVX512F.

func MaskPermutex2varEpi32 

func MaskPermutex2varEpi32(a x86.M128i, k x86.Mmask8, idx x86.M128i, b x86.M128i) (dst x86.M128i)

MaskPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+2] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMT2D'. Intrinsic: '_mm_mask_permutex2var_epi32'. Requires AVX512F.

func MaskPermutex2varEpi64 

func MaskPermutex2varEpi64(a x86.M128i, k x86.Mmask8, idx x86.M128i, b x86.M128i) (dst x86.M128i)

MaskPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMT2Q'. Intrinsic: '_mm_mask_permutex2var_epi64'. Requires AVX512F.

func MaskPermutex2varPd 

func MaskPermutex2varPd(a x86.M128d, k x86.Mmask8, idx x86.M128i, b x86.M128d) (dst x86.M128d)

MaskPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := a[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMT2PD'. Intrinsic: '_mm_mask_permutex2var_pd'. Requires AVX512F.

func MaskPermutex2varPs 

func MaskPermutex2varPs(a x86.M128, k x86.Mmask8, idx x86.M128i, b x86.M128) (dst x86.M128)

MaskPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using writemask 'k' (elements are copied from 'a' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := idx[i+2] ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := a[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMT2PS'. Intrinsic: '_mm_mask_permutex2var_ps'. Requires AVX512F.

func MaskRcp14Pd 

func MaskRcp14Pd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm_mask_rcp14_pd'. Requires AVX512F.

func MaskRcp14Ps 

func MaskRcp14Ps(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm_mask_rcp14_ps'. Requires AVX512F.

func MaskRcp14Sd 

func MaskRcp14Sd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskRcp14Sd: Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[63:0] := APPROXIMATE(1.0/b[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRCP14SD'. Intrinsic: '_mm_mask_rcp14_sd'. Requires AVX512F.

func MaskRcp14Ss 

func MaskRcp14Ss(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskRcp14Ss: Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[31:0] := APPROXIMATE(1.0/b[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRCP14SS'. Intrinsic: '_mm_mask_rcp14_ss'. Requires AVX512F.

func MaskRolEpi32 

func MaskRolEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm_mask_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRolEpi64 

func MaskRolEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm_mask_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRolvEpi32 

func MaskRolvEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm_mask_rolv_epi32'. Requires AVX512F.

func MaskRolvEpi64 

func MaskRolvEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm_mask_rolv_epi64'. Requires AVX512F.

func MaskRorEpi32 

func MaskRorEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm_mask_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRorEpi64 

func MaskRorEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm_mask_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRorvEpi32 

func MaskRorvEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm_mask_rorv_epi32'. Requires AVX512F.

func MaskRorvEpi64 

func MaskRorvEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm_mask_rorv_epi64'. Requires AVX512F.

func MaskRoundscalePd 

func MaskRoundscalePd(src x86.M128d, k x86.Mmask8, a x86.M128d, imm8 byte) (dst x86.M128d)

MaskRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm_mask_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRoundscalePs 

func MaskRoundscalePs(src x86.M128, k x86.Mmask8, a x86.M128, imm8 byte) (dst x86.M128)

MaskRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm_mask_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRoundscaleRoundSd 

func MaskRoundscaleRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte, rounding int) (dst x86.M128d)

MaskRoundscaleRoundSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		IF k[0]
			dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_mask_roundscale_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRoundscaleRoundSs 

func MaskRoundscaleRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte, rounding int) (dst x86.M128)

MaskRoundscaleRoundSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		IF k[0]
			dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_mask_roundscale_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRoundscaleSd 

func MaskRoundscaleSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.M128d)

MaskRoundscaleSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

IF k[0]
	dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_mask_roundscale_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRoundscaleSs 

func MaskRoundscaleSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.M128)

MaskRoundscaleSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

IF k[0]
	dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_mask_roundscale_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskRsqrt14Pd 

func MaskRsqrt14Pd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm_mask_rsqrt14_pd'. Requires AVX512F.

func MaskRsqrt14Ps 

func MaskRsqrt14Ps(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm_mask_rsqrt14_ps'. Requires AVX512F.

func MaskRsqrt14Sd 

func MaskRsqrt14Sd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskRsqrt14Sd: Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[63:0] := APPROXIMATE(1.0 / SQRT(b[63:0]))
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SD'. Intrinsic: '_mm_mask_rsqrt14_sd'. Requires AVX512F.

func MaskRsqrt14Ss 

func MaskRsqrt14Ss(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskRsqrt14Ss: Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[31:0] := APPROXIMATE(1.0 / SQRT(b[31:0]))
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SS'. Intrinsic: '_mm_mask_rsqrt14_ss'. Requires AVX512F.

func MaskScalefPd 

func MaskScalefPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm_mask_scalef_pd'. Requires AVX512F.

func MaskScalefPs 

func MaskScalefPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm_mask_scalef_ps'. Requires AVX512F.

func MaskScalefRoundSd 

func MaskScalefRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskScalefRoundSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		IF k[0]
			dst[63:0] := SCALE(a[63:0], b[63:0])
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_mask_scalef_round_sd'. Requires AVX512F.

func MaskScalefRoundSs 

func MaskScalefRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskScalefRoundSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[63:0]
		}

		IF k[0]
			dst[31:0] := SCALE(a[31:0], b[31:0])
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_mask_scalef_round_ss'. Requires AVX512F.

func MaskScalefSd 

func MaskScalefSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskScalefSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

IF k[0]
	dst[63:0] := SCALE(a[63:0], b[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_mask_scalef_sd'. Requires AVX512F.

func MaskScalefSs 

func MaskScalefSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskScalefSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[63:0]
}

IF k[0]
	dst[31:0] := SCALE(a[31:0], b[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_mask_scalef_ss'. Requires AVX512F.

func MaskSet1Epi32 

func MaskSet1Epi32(src x86.M128i, k x86.Mmask8, a int) (dst x86.M128i)

MaskSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm_mask_set1_epi32'. Requires AVX512F.

func MaskSet1Epi64 

func MaskSet1Epi64(src x86.M128i, k x86.Mmask8, a int64) (dst x86.M128i)

MaskSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm_mask_set1_epi64'. Requires AVX512F.

func MaskShuffleEpi32 

func MaskShuffleEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskShuffleEpi32: Shuffle 32-bit integers in 'a' using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSHUFD'. Intrinsic: '_mm_mask_shuffle_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskShufflePd 

func MaskShufflePd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.M128d)

MaskShufflePd: Shuffle double-precision (64-bit) floating-point elements using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm_mask_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskShufflePs 

func MaskShufflePs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.M128)

MaskShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm_mask_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSllEpi32 

func MaskSllEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm_mask_sll_epi32'. Requires AVX512F.

func MaskSllEpi64 

func MaskSllEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm_mask_sll_epi64'. Requires AVX512F.

func MaskSlliEpi32 

func MaskSlliEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSlliEpi32: Shift packed 32-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm_mask_slli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSlliEpi64 

func MaskSlliEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm_mask_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSllvEpi32 

func MaskSllvEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSllvEpi32: Shift packed 32-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLVD'. Intrinsic: '_mm_mask_sllv_epi32'. Requires AVX512F.

func MaskSllvEpi64 

func MaskSllvEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm_mask_sllv_epi64'. Requires AVX512F.

func MaskSqrtPd 

func MaskSqrtPd(src x86.M128d, k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm_mask_sqrt_pd'. Requires AVX512F.

func MaskSqrtPs 

func MaskSqrtPs(src x86.M128, k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm_mask_sqrt_ps'. Requires AVX512F.

func MaskSqrtRoundSd 

func MaskSqrtRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskSqrtRoundSd: Compute the square root of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := SQRT(a[63:0])
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSQRTSD'. Intrinsic: '_mm_mask_sqrt_round_sd'. Requires AVX512F.

func MaskSqrtRoundSs 

func MaskSqrtRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskSqrtRoundSs: Compute the square root of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := SQRT(a[31:0])
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSQRTSS'. Intrinsic: '_mm_mask_sqrt_round_ss'. Requires AVX512F.

func MaskSqrtSd 

func MaskSqrtSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskSqrtSd: Compute the square root of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := SQRT(a[63:0])
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VSQRTSD'. Intrinsic: '_mm_mask_sqrt_sd'. Requires AVX512F.

func MaskSqrtSs 

func MaskSqrtSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskSqrtSs: Compute the square root of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := SQRT(a[31:0])
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VSQRTSS'. Intrinsic: '_mm_mask_sqrt_ss'. Requires AVX512F.

func MaskSraEpi32 

func MaskSraEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm_mask_sra_epi32'. Requires AVX512F.

func MaskSraEpi64 

func MaskSraEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_mask_sra_epi64'. Requires AVX512F.

func MaskSraiEpi32 

func MaskSraiEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSraiEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm_mask_srai_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSraiEpi64 

func MaskSraiEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_mask_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSravEpi32 

func MaskSravEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSravEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAVD'. Intrinsic: '_mm_mask_srav_epi32'. Requires AVX512F.

func MaskSravEpi64 

func MaskSravEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm_mask_srav_epi64'. Requires AVX512F.

func MaskSrlEpi32 

func MaskSrlEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm_mask_srl_epi32'. Requires AVX512F.

func MaskSrlEpi64 

func MaskSrlEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm_mask_srl_epi64'. Requires AVX512F.

func MaskSrliEpi32 

func MaskSrliEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSrliEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm_mask_srli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSrliEpi64 

func MaskSrliEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm_mask_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskSrlvEpi32 

func MaskSrlvEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSrlvEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLVD'. Intrinsic: '_mm_mask_srlv_epi32'. Requires AVX512F.

func MaskSrlvEpi64 

func MaskSrlvEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm_mask_srlv_epi64'. Requires AVX512F.

func MaskStoreSd 

func MaskStoreSd(mem_addr *float64, k x86.Mmask8, a x86.M128d)

MaskStoreSd: Store the lower double-precision (64-bit) floating-point element from 'a' into memory using writemask 'k'. 

'mem_addr' must be aligned on a 16-byte boundary or a general-protection

exception may be generated. 

IF k[0]
	MEM[mem_addr+63:mem_addr] := a[63:0]
FI

Instruction: 'VMOVSD'. Intrinsic: '_mm_mask_store_sd'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskStoreSs 

func MaskStoreSs(mem_addr *float32, k x86.Mmask8, a x86.M128)

MaskStoreSs: Store the lower single-precision (32-bit) floating-point element from 'a' into memory using writemask 'k'. 

'mem_addr' must be aligned on a 16-byte boundary or a general-protection

exception may be generated. 

IF k[0]
	MEM[mem_addr+31:mem_addr] := a[31:0]
FI

Instruction: 'VMOVSS'. Intrinsic: '_mm_mask_store_ss'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskSubEpi32 

func MaskSubEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskSubEpi32: Subtract packed 32-bit integers in 'b' from packed 32-bit integers in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSUBD'. Intrinsic: '_mm_mask_sub_epi32'. Requires AVX512F.

func MaskSubEpi64 

func MaskSubEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm_mask_sub_epi64'. Requires AVX512F.

func MaskSubPd 

func MaskSubPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskSubPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm_mask_sub_pd'. Requires AVX512F.

func MaskSubPs 

func MaskSubPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskSubPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm_mask_sub_ps'. Requires AVX512F.

func MaskSubRoundSd 

func MaskSubRoundSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskSubRoundSd: Subtract the lower double-precision (64-bit) floating-point element in 'b' from the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] - b[63:0]
		ELSE
			dst[63:0] := src[63:0]
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VSUBSD'. Intrinsic: '_mm_mask_sub_round_sd'. Requires AVX512F.

func MaskSubRoundSs 

func MaskSubRoundSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskSubRoundSs: Subtract the lower single-precision (32-bit) floating-point element in 'b' from the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] - b[31:0]
		ELSE
			dst[31:0] := src[31:0]
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VSUBSS'. Intrinsic: '_mm_mask_sub_round_ss'. Requires AVX512F.

func MaskSubSd 

func MaskSubSd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskSubSd: Subtract the lower double-precision (64-bit) floating-point element in 'b' from the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] - b[63:0]
ELSE
	dst[63:0] := src[63:0]
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VSUBSD'. Intrinsic: '_mm_mask_sub_sd'. Requires AVX512F.

func MaskSubSs 

func MaskSubSs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskSubSs: Subtract the lower single-precision (32-bit) floating-point element in 'b' from the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using writemask 'k' (the element is copied from 'src' when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] - b[31:0]
ELSE
	dst[31:0] := src[31:0]
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VSUBSS'. Intrinsic: '_mm_mask_sub_ss'. Requires AVX512F.

func MaskTernarylogicEpi32 

func MaskTernarylogicEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.M128i)

MaskTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 32-bit granularity (32-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm_mask_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskTernarylogicEpi64 

func MaskTernarylogicEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i, imm8 byte) (dst x86.M128i)

MaskTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'src', 'a', and 'b' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using writemask 'k' at 64-bit granularity (64-bit elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (src[i+h] << 2) OR (a[i+h] << 1) OR b[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm_mask_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskTestEpi32Mask 

func MaskTestEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskTestEpi32Mask: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is non-zero. 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ((a[i+31:i] AND b[i+31:i]) != 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTMD'. Intrinsic: '_mm_mask_test_epi32_mask'. Requires AVX512F.

func MaskTestEpi64Mask 

func MaskTestEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskTestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is non-zero. 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm_mask_test_epi64_mask'. Requires AVX512F.

func MaskTestnEpi32Mask 

func MaskTestnEpi32Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskTestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 3
	i := j*32
	IF k1[j]
		k[j] := ((a[i+31:i] AND b[i+31:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm_mask_testn_epi32_mask'. Requires AVX512F.

func MaskTestnEpi64Mask 

func MaskTestnEpi64Mask(k1 x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.Mmask8)

MaskTestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' (subject to writemask 'k') if the intermediate value is zero. 

FOR j := 0 to 1
	i := j*64
	IF k1[j]
		k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
	ELSE
		k[j] := 0
	FI
ENDFOR
k[MAX:2] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm_mask_testn_epi64_mask'. Requires AVX512F.

func MaskUnpackhiEpi32 

func MaskUnpackhiEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm_mask_unpackhi_epi32'. Requires AVX512F.

func MaskUnpackhiEpi64 

func MaskUnpackhiEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm_mask_unpackhi_epi64'. Requires AVX512F.

func MaskUnpackhiPd 

func MaskUnpackhiPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm_mask_unpackhi_pd'. Requires AVX512F.

func MaskUnpackhiPs 

func MaskUnpackhiPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm_mask_unpackhi_ps'. Requires AVX512F.

func MaskUnpackloEpi32 

func MaskUnpackloEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm_mask_unpacklo_epi32'. Requires AVX512F.

func MaskUnpackloEpi64 

func MaskUnpackloEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm_mask_unpacklo_epi64'. Requires AVX512F.

func MaskUnpackloPd 

func MaskUnpackloPd(src x86.M128d, k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm_mask_unpacklo_pd'. Requires AVX512F.

func MaskUnpackloPs 

func MaskUnpackloPs(src x86.M128, k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm_mask_unpacklo_ps'. Requires AVX512F.

func MaskXorEpi32 

func MaskXorEpi32(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskXorEpi32: Compute the bitwise XOR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] XOR b[i+31:i]
	ELSE
		dst[i+31:i] := src[i+31:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPXORD'. Intrinsic: '_mm_mask_xor_epi32'. Requires AVX512F.

func MaskXorEpi64 

func MaskXorEpi64(src x86.M128i, k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskXorEpi64: Compute the bitwise XOR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using writemask 'k' (elements are copied from 'src' when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] XOR b[i+63:i]
	ELSE
		dst[i+63:i] := src[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm_mask_xor_epi64'. Requires AVX512F.

func MaskzAbsEpi32 

func MaskzAbsEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzAbsEpi32: Compute the absolute value of packed 32-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ABS(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPABSD'. Intrinsic: '_mm_maskz_abs_epi32'. Requires AVX512F.

func MaskzAbsEpi64 

func MaskzAbsEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzAbsEpi64: Compute the absolute value of packed 64-bit integers in 'a', and store the unsigned results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ABS(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPABSQ'. Intrinsic: '_mm_maskz_abs_epi64'. Requires AVX512F.

func MaskzAddEpi32 

func MaskzAddEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAddEpi32: Add packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] + b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPADDD'. Intrinsic: '_mm_maskz_add_epi32'. Requires AVX512F.

func MaskzAddEpi64 

func MaskzAddEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAddEpi64: Add packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] + b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPADDQ'. Intrinsic: '_mm_maskz_add_epi64'. Requires AVX512F.

func MaskzAddRoundSd 

func MaskzAddRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzAddRoundSd: Add the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] + b[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VADDSD'. Intrinsic: '_mm_maskz_add_round_sd'. Requires AVX512F.

func MaskzAddRoundSs 

func MaskzAddRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzAddRoundSs: Add the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] + b[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VADDSS'. Intrinsic: '_mm_maskz_add_round_ss'. Requires AVX512F.

func MaskzAddSd 

func MaskzAddSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzAddSd: Add the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] + b[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VADDSD'. Intrinsic: '_mm_maskz_add_sd'. Requires AVX512F.

func MaskzAddSs 

func MaskzAddSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzAddSs: Add the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] + b[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VADDSS'. Intrinsic: '_mm_maskz_add_ss'. Requires AVX512F.

func MaskzAndEpi32 

func MaskzAndEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAndEpi32: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDD'. Intrinsic: '_mm_maskz_and_epi32'. Requires AVX512F.

func MaskzAndEpi64 

func MaskzAndEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAndEpi64: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDQ'. Intrinsic: '_mm_maskz_and_epi64'. Requires AVX512F.

func MaskzAndnotEpi32 

func MaskzAndnotEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAndnotEpi32: Compute the bitwise AND NOT of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (NOT a[i+31:i]) AND b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDND'. Intrinsic: '_mm_maskz_andnot_epi32'. Requires AVX512F.

func MaskzAndnotEpi64 

func MaskzAndnotEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzAndnotEpi64: Compute the bitwise AND NOT of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (NOT a[i+63:i]) AND b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPANDNQ'. Intrinsic: '_mm_maskz_andnot_epi64'. Requires AVX512F.

func MaskzBroadcastdEpi32 

func MaskzBroadcastdEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzBroadcastdEpi32: Broadcast the low packed 32-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm_maskz_broadcastd_epi32'. Requires AVX512F.

func MaskzBroadcastqEpi64 

func MaskzBroadcastqEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzBroadcastqEpi64: Broadcast the low packed 64-bit integer from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm_maskz_broadcastq_epi64'. Requires AVX512F.

func MaskzBroadcastssPs 

func MaskzBroadcastssPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzBroadcastssPs: Broadcast the low single-precision (32-bit) floating-point element from 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VBROADCASTSS'. Intrinsic: '_mm_maskz_broadcastss_ps'. Requires AVX512F.

func MaskzCompressEpi32 

func MaskzCompressEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCompressEpi32: Contiguously store the active 32-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := 0
dst[MAX:128] := 0

Instruction: 'VPCOMPRESSD'. Intrinsic: '_mm_maskz_compress_epi32'. Requires AVX512F.

func MaskzCompressEpi64 

func MaskzCompressEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCompressEpi64: Contiguously store the active 64-bit integers in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := 0
dst[MAX:128] := 0

Instruction: 'VPCOMPRESSQ'. Intrinsic: '_mm_maskz_compress_epi64'. Requires AVX512F.

func MaskzCompressPd 

func MaskzCompressPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzCompressPd: Contiguously store the active double-precision (64-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 64
m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[m+size-1:m] := a[i+63:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := 0
dst[MAX:128] := 0

Instruction: 'VCOMPRESSPD'. Intrinsic: '_mm_maskz_compress_pd'. Requires AVX512F.

func MaskzCompressPs 

func MaskzCompressPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzCompressPs: Contiguously store the active single-precision (32-bit) floating-point elements in 'a' (those with their respective bit set in zeromask 'k') to 'dst', and set the remaining elements to zero. 

size := 32
m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[m+size-1:m] := a[i+31:i]
		m := m + size
	FI
ENDFOR
dst[127:m] := 0
dst[MAX:128] := 0

Instruction: 'VCOMPRESSPS'. Intrinsic: '_mm_maskz_compress_ps'. Requires AVX512F.

func MaskzCvtRoundpsPh 

func MaskzCvtRoundpsPh(k x86.Mmask8, a x86.M128, rounding int) (dst x86.M128i)

MaskzCvtRoundpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 3
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:64] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm_maskz_cvt_roundps_ph'. Requires AVX512F.

func MaskzCvtRoundsdSs 

func MaskzCvtRoundsdSs(k x86.Mmask8, a x86.M128, b x86.M128d, rounding int) (dst x86.M128)

MaskzCvtRoundsdSs: Convert the lower double-precision (64-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := Convert_FP64_To_FP32(b[63:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:31]
		dst[MAX:64] := 0

Instruction: 'VCVTSD2SS'. Intrinsic: '_mm_maskz_cvt_roundsd_ss'. Requires AVX512F.

func MaskzCvtRoundssSd 

func MaskzCvtRoundssSd(k x86.Mmask8, a x86.M128d, b x86.M128, rounding int) (dst x86.M128d)

MaskzCvtRoundssSd: Convert the lower single-precision (32-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := Convert_FP32_To_FP64(b[31:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:64] := 0

Instruction: 'VCVTSS2SD'. Intrinsic: '_mm_maskz_cvt_roundss_sd'. Requires AVX512F.

func MaskzCvtepi16Epi32 

func MaskzCvtepi16Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi16Epi32: Sign extend packed 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXWD'. Intrinsic: '_mm_maskz_cvtepi16_epi32'. Requires AVX512F.

func MaskzCvtepi16Epi64 

func MaskzCvtepi16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi16Epi64: Sign extend packed 16-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXWQ'. Intrinsic: '_mm_maskz_cvtepi16_epi64'. Requires AVX512F.

func MaskzCvtepi32Epi16 

func MaskzCvtepi32Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVDW'. Intrinsic: '_mm_maskz_cvtepi32_epi16'. Requires AVX512F.

func MaskzCvtepi32Epi64 

func MaskzCvtepi32Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi32Epi64: Sign extend packed 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXDQ'. Intrinsic: '_mm_maskz_cvtepi32_epi64'. Requires AVX512F.

func MaskzCvtepi32Epi8 

func MaskzCvtepi32Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVDB'. Intrinsic: '_mm_maskz_cvtepi32_epi8'. Requires AVX512F.

func MaskzCvtepi32Pd 

func MaskzCvtepi32Pd(k x86.Mmask8, a x86.M128i) (dst x86.M128d)

MaskzCvtepi32Pd: Convert packed 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*32
	m := j*64
	IF k[j]
		dst[m+63:m] := Convert_Int32_To_FP64(a[i+31:i])
	ELSE
		dst[m+63:m] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTDQ2PD'. Intrinsic: '_mm_maskz_cvtepi32_pd'. Requires AVX512F.

func MaskzCvtepi32Ps 

func MaskzCvtepi32Ps(k x86.Mmask8, a x86.M128i) (dst x86.M128)

MaskzCvtepi32Ps: Convert packed 32-bit integers in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_Int32_To_FP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTDQ2PS'. Intrinsic: '_mm_maskz_cvtepi32_ps'. Requires AVX512F.

func MaskzCvtepi64Epi16 

func MaskzCvtepi64Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Truncate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVQW'. Intrinsic: '_mm_maskz_cvtepi64_epi16'. Requires AVX512F.

func MaskzCvtepi64Epi32 

func MaskzCvtepi64Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Truncate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVQD'. Intrinsic: '_mm_maskz_cvtepi64_epi32'. Requires AVX512F.

func MaskzCvtepi64Epi8 

func MaskzCvtepi64Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Truncate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVQB'. Intrinsic: '_mm_maskz_cvtepi64_epi8'. Requires AVX512F.

func MaskzCvtepi8Epi32 

func MaskzCvtepi8Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi8Epi32: Sign extend packed 8-bit integers in the low 4 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXBD'. Intrinsic: '_mm_maskz_cvtepi8_epi32'. Requires AVX512F.

func MaskzCvtepi8Epi64 

func MaskzCvtepi8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepi8Epi64: Sign extend packed 8-bit integers in the low 2 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := SignExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVSXBQ'. Intrinsic: '_mm_maskz_cvtepi8_epi64'. Requires AVX512F.

func MaskzCvtepu16Epi32 

func MaskzCvtepu16Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepu16Epi32: Zero extend packed unsigned 16-bit integers in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXWD'. Intrinsic: '_mm_maskz_cvtepu16_epi32'. Requires AVX512F.

func MaskzCvtepu16Epi64 

func MaskzCvtepu16Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepu16Epi64: Zero extend packed unsigned 16-bit integers in the low 4 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+15:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXWQ'. Intrinsic: '_mm_maskz_cvtepu16_epi64'. Requires AVX512F.

func MaskzCvtepu32Epi64 

func MaskzCvtepu32Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepu32Epi64: Zero extend packed unsigned 32-bit integers in 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXDQ'. Intrinsic: '_mm_maskz_cvtepu32_epi64'. Requires AVX512F.

func MaskzCvtepu32Pd 

func MaskzCvtepu32Pd(k x86.Mmask8, a x86.M128i) (dst x86.M128d)

MaskzCvtepu32Pd: Convert packed unsigned 32-bit integers in 'a' to packed double-precision (64-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	l := j*32
	IF k[j]
		dst[i+63:i] := ConvertUnsignedIntegerTo_FP64(a[l+31:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTUDQ2PD'. Intrinsic: '_mm_maskz_cvtepu32_pd'. Requires AVX512F.

func MaskzCvtepu8Epi32 

func MaskzCvtepu8Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepu8Epi32: Zero extend packed unsigned 8-bit integers in th elow 4 bytes of 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXBD'. Intrinsic: '_mm_maskz_cvtepu8_epi32'. Requires AVX512F.

func MaskzCvtepu8Epi64 

func MaskzCvtepu8Epi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtepu8Epi64: Zero extend packed unsigned 8-bit integers in the low 2 bytes of 'a' to packed 64-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[l+7:l])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMOVZXBQ'. Intrinsic: '_mm_maskz_cvtepu8_epi64'. Requires AVX512F.

func MaskzCvtpdEpi32 

func MaskzCvtpdEpi32(k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskzCvtpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2DQ'. Intrinsic: '_mm_maskz_cvtpd_epi32'. Requires AVX512F.

func MaskzCvtpdEpu32 

func MaskzCvtpdEpu32(k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskzCvtpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2UDQ'. Intrinsic: '_mm_maskz_cvtpd_epu32'. Requires AVX512F.

func MaskzCvtpdPs 

func MaskzCvtpdPs(k x86.Mmask8, a x86.M128d) (dst x86.M128)

MaskzCvtpdPs: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*32
	l := j*64
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_FP32(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTPD2PS'. Intrinsic: '_mm_maskz_cvtpd_ps'. Requires AVX512F.

func MaskzCvtphPs 

func MaskzCvtphPs(k x86.Mmask8, a x86.M128i) (dst x86.M128)

MaskzCvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	m := j*16
	IF k[j]
		dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm_maskz_cvtph_ps'. Requires AVX512F.

func MaskzCvtpsEpi32 

func MaskzCvtpsEpi32(k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskzCvtpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_Int32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPS2DQ'. Intrinsic: '_mm_maskz_cvtps_epi32'. Requires AVX512F.

func MaskzCvtpsEpu32 

func MaskzCvtpsEpu32(k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskzCvtpsEpu32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_UnsignedInt32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPS2UDQ'. Intrinsic: '_mm_maskz_cvtps_epu32'. Requires AVX512F.

func MaskzCvtpsPh 

func MaskzCvtpsPh(k x86.Mmask8, a x86.M128, rounding int) (dst x86.M128i)

MaskzCvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 3
			i := 16*j
			l := 32*j
			IF k[j]
				dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
			ELSE
				dst[i+15:i] := 0
			FI
		ENDFOR
		dst[MAX:64] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm_maskz_cvtps_ph'. Requires AVX512F.

func MaskzCvtsdSs 

func MaskzCvtsdSs(k x86.Mmask8, a x86.M128, b x86.M128d) (dst x86.M128)

MaskzCvtsdSs: Convert the lower double-precision (64-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point element, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := Convert_FP64_To_FP32(b[63:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:31]
dst[MAX:64] := 0

Instruction: 'VCVTSD2SS'. Intrinsic: '_mm_maskz_cvtsd_ss'. Requires AVX512F.

func MaskzCvtsepi32Epi16 

func MaskzCvtsepi32Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtsepi32Epi16: Convert packed 32-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSDW'. Intrinsic: '_mm_maskz_cvtsepi32_epi16'. Requires AVX512F.

func MaskzCvtsepi32Epi8 

func MaskzCvtsepi32Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtsepi32Epi8: Convert packed 32-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSDB'. Intrinsic: '_mm_maskz_cvtsepi32_epi8'. Requires AVX512F.

func MaskzCvtsepi64Epi16 

func MaskzCvtsepi64Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtsepi64Epi16: Convert packed 64-bit integers in 'a' to packed 16-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_Int64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVSQW'. Intrinsic: '_mm_maskz_cvtsepi64_epi16'. Requires AVX512F.

func MaskzCvtsepi64Epi32 

func MaskzCvtsepi64Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtsepi64Epi32: Convert packed 64-bit integers in 'a' to packed 32-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_Int64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVSQD'. Intrinsic: '_mm_maskz_cvtsepi64_epi32'. Requires AVX512F.

func MaskzCvtsepi64Epi8 

func MaskzCvtsepi64Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtsepi64Epi8: Convert packed 64-bit integers in 'a' to packed 8-bit integers with signed saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_Int64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVSQB'. Intrinsic: '_mm_maskz_cvtsepi64_epi8'. Requires AVX512F.

func MaskzCvtssSd 

func MaskzCvtssSd(k x86.Mmask8, a x86.M128d, b x86.M128) (dst x86.M128d)

MaskzCvtssSd: Convert the lower single-precision (32-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point element, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := Convert_FP32_To_FP64(b[31:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:64] := 0

Instruction: 'VCVTSS2SD'. Intrinsic: '_mm_maskz_cvtss_sd'. Requires AVX512F.

func MaskzCvttpdEpi32 

func MaskzCvttpdEpi32(k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskzCvttpdEpi32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_Int32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTTPD2DQ'. Intrinsic: '_mm_maskz_cvttpd_epi32'. Requires AVX512F.

func MaskzCvttpdEpu32 

func MaskzCvttpdEpu32(k x86.Mmask8, a x86.M128d) (dst x86.M128i)

MaskzCvttpdEpu32: Convert packed double-precision (64-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 32*j
	l := 64*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[l+63:l])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VCVTTPD2UDQ'. Intrinsic: '_mm_maskz_cvttpd_epu32'. Requires AVX512F.

func MaskzCvttpsEpi32 

func MaskzCvttpsEpi32(k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskzCvttpsEpi32: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*i
	IF k[j]
		dst[i+31:i] := Convert_FP32_To_IntegerTruncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPS2DQ'. Intrinsic: '_mm_maskz_cvttps_epi32'. Requires AVX512F.

func MaskzCvttpsEpu32 

func MaskzCvttpsEpu32(k x86.Mmask8, a x86.M128) (dst x86.M128i)

MaskzCvttpsEpu32: Convert packed double-precision (32-bit) floating-point elements in 'a' to packed unsigned 32-bit integers with truncation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := Convert_FP64_To_UnsignedInt32_Truncate(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTTPS2UDQ'. Intrinsic: '_mm_maskz_cvttps_epu32'. Requires AVX512F.

func MaskzCvtusepi32Epi16 

func MaskzCvtusepi32Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtusepi32Epi16: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt32_To_Int16(a[i+31:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSDW'. Intrinsic: '_mm_maskz_cvtusepi32_epi16'. Requires AVX512F.

func MaskzCvtusepi32Epi8 

func MaskzCvtusepi32Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtusepi32Epi8: Convert packed unsigned 32-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt32_To_Int8(a[i+31:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSDB'. Intrinsic: '_mm_maskz_cvtusepi32_epi8'. Requires AVX512F.

func MaskzCvtusepi64Epi16 

func MaskzCvtusepi64Epi16(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtusepi64Epi16: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 16-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 16*j
	IF k[j]
		dst[l+15:l] := Saturate_UnsignedInt64_To_Int16(a[i+63:i])
	ELSE
		dst[l+15:l] := 0
	FI
ENDFOR
dst[MAX:32] := 0

Instruction: 'VPMOVUSQW'. Intrinsic: '_mm_maskz_cvtusepi64_epi16'. Requires AVX512F.

func MaskzCvtusepi64Epi32 

func MaskzCvtusepi64Epi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtusepi64Epi32: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 32-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 32*j
	IF k[j]
		dst[l+31:l] := Saturate_UnsignedInt64_To_Int32(a[i+63:i])
	ELSE
		dst[l+31:l] := 0
	FI
ENDFOR
dst[MAX:64] := 0

Instruction: 'VPMOVUSQD'. Intrinsic: '_mm_maskz_cvtusepi64_epi32'. Requires AVX512F.

func MaskzCvtusepi64Epi8 

func MaskzCvtusepi64Epi8(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzCvtusepi64Epi8: Convert packed unsigned 64-bit integers in 'a' to packed unsigned 8-bit integers with unsigned saturation, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	l := 8*j
	IF k[j]
		dst[l+7:l] := Saturate_UnsignedInt64_To_Int8(a[i+63:i])
	ELSE
		dst[l+7:l] := 0
	FI
ENDFOR
dst[MAX:16] := 0

Instruction: 'VPMOVUSQB'. Intrinsic: '_mm_maskz_cvtusepi64_epi8'. Requires AVX512F.

func MaskzDivPd 

func MaskzDivPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzDivPd: Divide packed double-precision (64-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := 64*j
	IF k[j]
		dst[i+63:i] := a[i+63:i] / b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VDIVPD'. Intrinsic: '_mm_maskz_div_pd'. Requires AVX512F.

func MaskzDivPs 

func MaskzDivPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzDivPs: Divide packed single-precision (32-bit) floating-point elements in 'a' by packed elements in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := 32*j
	IF k[j]
		dst[i+31:i] := a[i+31:i] / b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VDIVPS'. Intrinsic: '_mm_maskz_div_ps'. Requires AVX512F.

func MaskzDivRoundSd 

func MaskzDivRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzDivRoundSd: Divide the lower double-precision (64-bit) floating-point element in 'a' by the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] / b[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VDIVSD'. Intrinsic: '_mm_maskz_div_round_sd'. Requires AVX512F.

func MaskzDivRoundSs 

func MaskzDivRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzDivRoundSs: Divide the lower single-precision (32-bit) floating-point element in 'a' by the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] / b[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VDIVSS'. Intrinsic: '_mm_maskz_div_round_ss'. Requires AVX512F.

func MaskzDivSd 

func MaskzDivSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzDivSd: Divide the lower double-precision (64-bit) floating-point element in 'a' by the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] / b[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VDIVSD'. Intrinsic: '_mm_maskz_div_sd'. Requires AVX512F.

func MaskzDivSs 

func MaskzDivSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzDivSs: Divide the lower single-precision (32-bit) floating-point element in 'a' by the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] / b[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VDIVSS'. Intrinsic: '_mm_maskz_div_ss'. Requires AVX512F.

func MaskzExpandEpi32 

func MaskzExpandEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzExpandEpi32: Load contiguous active 32-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPEXPANDD'. Intrinsic: '_mm_maskz_expand_epi32'. Requires AVX512F.

func MaskzExpandEpi64 

func MaskzExpandEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzExpandEpi64: Load contiguous active 64-bit integers from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPEXPANDQ'. Intrinsic: '_mm_maskz_expand_epi64'. Requires AVX512F.

func MaskzExpandPd 

func MaskzExpandPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzExpandPd: Load contiguous active double-precision (64-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[m+63:m]
		m := m + 64
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXPANDPD'. Intrinsic: '_mm_maskz_expand_pd'. Requires AVX512F.

func MaskzExpandPs 

func MaskzExpandPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzExpandPs: Load contiguous active single-precision (32-bit) floating-point elements from 'a' (those with their respective bit set in mask 'k'), and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

m := 0
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[m+31:m]
		m := m + 32
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VEXPANDPS'. Intrinsic: '_mm_maskz_expand_ps'. Requires AVX512F.

func MaskzFixupimmPd 

func MaskzFixupimmPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

MaskzFixupimmPd: Fix up packed double-precision (64-bit) floating-point elements in 'a' and 'b' using packed 64-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := FIXUPIMMPD(a[i+63:i], b[i+63:i], c[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPD'. Intrinsic: '_mm_maskz_fixupimm_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFixupimmPs 

func MaskzFixupimmPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

MaskzFixupimmPs: Fix up packed single-precision (32-bit) floating-point elements in 'a' and 'b' using packed 32-bit integers in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := FIXUPIMMPD(a[i+31:i], b[i+31:i], c[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMPS'. Intrinsic: '_mm_maskz_fixupimm_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFixupimmRoundSd 

func MaskzFixupimmRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte, rounding int) (dst x86.M128d)

MaskzFixupimmRoundSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN := 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
			tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
			CASE(tsrc[63:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[63:0] := src1[63:0]
			1 : dest[63:0] := tsrc[63:0]
			2 : dest[63:0] := QNaN(tsrc[63:0])
			3 : dest[63:0] := QNAN_Indefinite
			4 : dest[63:0] := -INF
			5 : dest[63:0] := +INF
			6 : dest[63:0] := tsrc.sign? –INF : +INF
			7 : dest[63:0] := -0
			8 : dest[63:0] := +0
			9 : dest[63:0] := -1
			10: dest[63:0] := +1
			11: dest[63:0] := 1⁄2
			12: dest[63:0] := 90.0
			13: dest[63:0] := PI/2
			14: dest[63:0] := MAX_FLOAT
			15: dest[63:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[63:0]
		}

		IF k[0]
			dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_maskz_fixupimm_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFixupimmRoundSs 

func MaskzFixupimmRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128i, imm8 byte, rounding int) (dst x86.M128)

MaskzFixupimmRoundSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		enum TOKEN_TYPE {
			QNAN_TOKEN := 0,
			SNAN_TOKEN L= 1,
			ZERO_VALUE_TOKEN := 2,
			ONE_VALUE_TOKEN := 3,
			NEG_INF_TOKEN := 4,
			POS_INF_TOKEN := 5,
			NEG_VALUE_TOKEN := 6,
			POS_VALUE_TOKEN := 7
		}
		FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
			tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
			CASE(tsrc[31:0] of TOKEN_TYPE)
			QNAN_TOKEN:j := 0
			SNAN_TOKEN:j := 1
			ZERO_VALUE_TOKEN: j := 2
			ONE_VALUE_TOKEN: j := 3
			NEG_INF_TOKEN: j := 4
			POS_INF_TOKEN: j := 5
			NEG_VALUE_TOKEN: j := 6
			POS_VALUE_TOKEN: j := 7
			ESAC

			token_response[3:0] := src3[3+4*j:4*j]

			CASE(token_response[3:0]) of
			0 : dest[31:0] := src1[31:0]
			1 : dest[31:0] := tsrc[31:0]
			2 : dest[31:0] := QNaN(tsrc[31:0])
			3 : dest[31:0] := QNAN_Indefinite
			4 : dest[31:0] := -INF
			5 : dest[31:0] := +INF
			6 : dest[31:0] := tsrc.sign? –INF : +INF
			7 : dest[31:0] := -0
			8 : dest[31:0] := +0
			9 : dest[31:0] := -1
			10: dest[31:0] := +1
			11: dest[31:0] := 1⁄2
			12: dest[31:0] := 90.0
			13: dest[31:0] := PI/2
			14: dest[31:0] := MAX_FLOAT
			15: dest[31:0] := -MAX_FLOAT
			ESAC

			CASE(tsrc[31:0] of TOKEN_TYPE)
			ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
			ZERO_VALUE_TOKEN: if imm8[1] then set #IE
			ONE_VALUE_TOKEN: if imm8[2] then set #ZE
			ONE_VALUE_TOKEN: if imm8[3] then set #IE
			SNAN_TOKEN: if imm8[4] then set #IE
			NEG_INF_TOKEN: if imm8[5] then set #IE
			NEG_VALUE_TOKEN: if imm8[6] then set #IE
			POS_INF_TOKEN: if imm8[7] then set #IE
			ESAC
			RETURN dest[31:0]
		}

		IF k[0]
			dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_maskz_fixupimm_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFixupimmSd 

func MaskzFixupimmSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128i, imm8 byte) (dst x86.M128d)

MaskzFixupimmSd: Fix up the lower double-precision (64-bit) floating-point elements in 'a' and 'b' using the lower 64-bit integer in 'c', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN := 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[63:0], src2[63:0], src3[63:0], imm8[7:0]){
	tsrc[63:0] := ((src2[62:52] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[63:0]
	CASE(tsrc[63:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[63:0] := src1[63:0]
	1 : dest[63:0] := tsrc[63:0]
	2 : dest[63:0] := QNaN(tsrc[63:0])
	3 : dest[63:0] := QNAN_Indefinite
	4 : dest[63:0] := -INF
	5 : dest[63:0] := +INF
	6 : dest[63:0] := tsrc.sign? –INF : +INF
	7 : dest[63:0] := -0
	8 : dest[63:0] := +0
	9 : dest[63:0] := -1
	10: dest[63:0] := +1
	11: dest[63:0] := 1⁄2
	12: dest[63:0] := 90.0
	13: dest[63:0] := PI/2
	14: dest[63:0] := MAX_FLOAT
	15: dest[63:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[63:0]
}

IF k[0]
	dst[63:0] := FIXUPIMMPD(a[63:0], b[63:0], c[63:0], imm8[7:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSD'. Intrinsic: '_mm_maskz_fixupimm_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFixupimmSs 

func MaskzFixupimmSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128i, imm8 byte) (dst x86.M128)

MaskzFixupimmSs: Fix up the lower single-precision (32-bit) floating-point elements in 'a' and 'b' using the lower 32-bit integer in 'c', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 'imm8' is used to set the required flags reporting. 

enum TOKEN_TYPE {
	QNAN_TOKEN := 0,
	SNAN_TOKEN L= 1,
	ZERO_VALUE_TOKEN := 2,
	ONE_VALUE_TOKEN := 3,
	NEG_INF_TOKEN := 4,
	POS_INF_TOKEN := 5,
	NEG_VALUE_TOKEN := 6,
	POS_VALUE_TOKEN := 7
}
FIXUPIMMPD(src1[31:0], src2[31:0], src3[31:0], imm8[7:0]){
	tsrc[31:0] := ((src2[30:23] == 0) AND (MXCSR.DAZ == 1)) ? 0.0 : src2[31:0]
	CASE(tsrc[31:0] of TOKEN_TYPE)
	QNAN_TOKEN:j := 0
	SNAN_TOKEN:j := 1
	ZERO_VALUE_TOKEN: j := 2
	ONE_VALUE_TOKEN: j := 3
	NEG_INF_TOKEN: j := 4
	POS_INF_TOKEN: j := 5
	NEG_VALUE_TOKEN: j := 6
	POS_VALUE_TOKEN: j := 7
	ESAC

	token_response[3:0] := src3[3+4*j:4*j]

	CASE(token_response[3:0]) of
	0 : dest[31:0] := src1[31:0]
	1 : dest[31:0] := tsrc[31:0]
	2 : dest[31:0] := QNaN(tsrc[31:0])
	3 : dest[31:0] := QNAN_Indefinite
	4 : dest[31:0] := -INF
	5 : dest[31:0] := +INF
	6 : dest[31:0] := tsrc.sign? –INF : +INF
	7 : dest[31:0] := -0
	8 : dest[31:0] := +0
	9 : dest[31:0] := -1
	10: dest[31:0] := +1
	11: dest[31:0] := 1⁄2
	12: dest[31:0] := 90.0
	13: dest[31:0] := PI/2
	14: dest[31:0] := MAX_FLOAT
	15: dest[31:0] := -MAX_FLOAT
	ESAC

	CASE(tsrc[31:0] of TOKEN_TYPE)
	ZERO_VALUE_TOKEN: if imm8[0] then set #ZE
	ZERO_VALUE_TOKEN: if imm8[1] then set #IE
	ONE_VALUE_TOKEN: if imm8[2] then set #ZE
	ONE_VALUE_TOKEN: if imm8[3] then set #IE
	SNAN_TOKEN: if imm8[4] then set #IE
	NEG_INF_TOKEN: if imm8[5] then set #IE
	NEG_VALUE_TOKEN: if imm8[6] then set #IE
	POS_INF_TOKEN: if imm8[7] then set #IE
	ESAC
	RETURN dest[31:0]
}

IF k[0]
	dst[31:0] := FIXUPIMMPD(a[31:0], b[31:0], c[31:0], imm8[7:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFIXUPIMMSS'. Intrinsic: '_mm_maskz_fixupimm_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzFmaddPd 

func MaskzFmaddPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PD, VFMADD213PD, VFMADD231PD'. Intrinsic: '_mm_maskz_fmadd_pd'. Requires AVX512F.

func MaskzFmaddPs 

func MaskzFmaddPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADD132PS, VFMADD213PS, VFMADD231PS'. Intrinsic: '_mm_maskz_fmadd_ps'. Requires AVX512F.

func MaskzFmaddRoundSd 

func MaskzFmaddRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskzFmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_maskz_fmadd_round_sd'. Requires AVX512F.

func MaskzFmaddRoundSs 

func MaskzFmaddRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskzFmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_maskz_fmadd_round_ss'. Requires AVX512F.

func MaskzFmaddSd 

func MaskzFmaddSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMADD132SD, VFMADD213SD, VFMADD231SD'. Intrinsic: '_mm_maskz_fmadd_sd'. Requires AVX512F.

func MaskzFmaddSs 

func MaskzFmaddSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMADD132SS, VFMADD213SS, VFMADD231SS'. Intrinsic: '_mm_maskz_fmadd_ss'. Requires AVX512F.

func MaskzFmaddsubPd 

func MaskzFmaddsubPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmaddsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PD, VFMADDSUB213PD, VFMADDSUB231PD'. Intrinsic: '_mm_maskz_fmaddsub_pd'. Requires AVX512F.

func MaskzFmaddsubPs 

func MaskzFmaddsubPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmaddsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively add and subtract packed elements in 'c' to/from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMADDSUB132PS, VFMADDSUB213PS, VFMADDSUB231PS'. Intrinsic: '_mm_maskz_fmaddsub_ps'. Requires AVX512F.

func MaskzFmsubPd 

func MaskzFmsubPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PD, VFMSUB213PD, VFMSUB231PD'. Intrinsic: '_mm_maskz_fmsub_pd'. Requires AVX512F.

func MaskzFmsubPs 

func MaskzFmsubPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUB132PS, VFMSUB213PS, VFMSUB231PS'. Intrinsic: '_mm_maskz_fmsub_ps'. Requires AVX512F.

func MaskzFmsubRoundSd 

func MaskzFmsubRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskzFmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_maskz_fmsub_round_sd'. Requires AVX512F.

func MaskzFmsubRoundSs 

func MaskzFmsubRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskzFmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_maskz_fmsub_round_ss'. Requires AVX512F.

func MaskzFmsubSd 

func MaskzFmsubSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := (a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SD, VFMSUB213SD, VFMSUB231SD'. Intrinsic: '_mm_maskz_fmsub_sd'. Requires AVX512F.

func MaskzFmsubSs 

func MaskzFmsubSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := (a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFMSUB132SS, VFMSUB213SS, VFMSUB231SS'. Intrinsic: '_mm_maskz_fmsub_ss'. Requires AVX512F.

func MaskzFmsubaddPd 

func MaskzFmsubaddPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFmsubaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF (j is even)
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) + c[i+63:i]
		ELSE
			dst[i+63:i] := (a[i+63:i] * b[i+63:i]) - c[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PD, VFMSUBADD213PD, VFMSUBADD231PD'. Intrinsic: '_mm_maskz_fmsubadd_pd'. Requires AVX512F.

func MaskzFmsubaddPs 

func MaskzFmsubaddPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFmsubaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', alternatively subtract and add packed elements in 'c' from/to the intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF (j is even)
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) + c[i+31:i]
		ELSE
			dst[i+31:i] := (a[i+31:i] * b[i+31:i]) - c[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFMSUBADD132PS, VFMSUBADD213PS, VFMSUBADD231PS'. Intrinsic: '_mm_maskz_fmsubadd_ps'. Requires AVX512F.

func MaskzFnmaddPd 

func MaskzFnmaddPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFnmaddPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) + c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PD, VFNMADD213PD, VFNMADD231PD'. Intrinsic: '_mm_maskz_fnmadd_pd'. Requires AVX512F.

func MaskzFnmaddPs 

func MaskzFnmaddPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFnmaddPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', add the negated intermediate result to packed elements in 'c', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) + c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMADD132PS, VFNMADD213PS, VFNMADD231PS'. Intrinsic: '_mm_maskz_fnmadd_ps'. Requires AVX512F.

func MaskzFnmaddRoundSd 

func MaskzFnmaddRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskzFnmaddRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SD, VFNMADD213SD, VFNMADD231SD'. Intrinsic: '_mm_maskz_fnmadd_round_sd'. Requires AVX512F.

func MaskzFnmaddRoundSs 

func MaskzFnmaddRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskzFnmaddRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_maskz_fnmadd_round_ss'. Requires AVX512F.

func MaskzFnmaddSd 

func MaskzFnmaddSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFnmaddSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) + c[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMADD213SD, VFNMADD231SD, VFNMADD132SD'. Intrinsic: '_mm_maskz_fnmadd_sd'. Requires AVX512F.

func MaskzFnmaddSs 

func MaskzFnmaddSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFnmaddSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and add the negated intermediate result to the lower element in 'c'. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) + c[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMADD132SS, VFNMADD213SS, VFNMADD231SS'. Intrinsic: '_mm_maskz_fnmadd_ss'. Requires AVX512F.

func MaskzFnmsubPd 

func MaskzFnmsubPd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFnmsubPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := -(a[i+63:i] * b[i+63:i]) - c[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PD, VFNMSUB213PD, VFNMSUB231PD'. Intrinsic: '_mm_maskz_fnmsub_pd'. Requires AVX512F.

func MaskzFnmsubPs 

func MaskzFnmsubPs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFnmsubPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', subtract packed elements in 'c' from the negated intermediate result, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := -(a[i+31:i] * b[i+31:i]) - c[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VFNMSUB132PS, VFNMSUB213PS, VFNMSUB231PS'. Intrinsic: '_mm_maskz_fnmsub_ps'. Requires AVX512F.

func MaskzFnmsubRoundSd 

func MaskzFnmsubRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d, rounding int) (dst x86.M128d)

MaskzFnmsubRoundSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_maskz_fnmsub_round_sd'. Requires AVX512F.

func MaskzFnmsubRoundSs 

func MaskzFnmsubRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128, rounding int) (dst x86.M128)

MaskzFnmsubRoundSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_maskz_fnmsub_round_ss'. Requires AVX512F.

func MaskzFnmsubSd 

func MaskzFnmsubSd(k x86.Mmask8, a x86.M128d, b x86.M128d, c x86.M128d) (dst x86.M128d)

MaskzFnmsubSd: Multiply the lower double-precision (64-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := -(a[63:0] * b[63:0]) - c[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SD, VFNMSUB213SD, VFNMSUB231SD'. Intrinsic: '_mm_maskz_fnmsub_sd'. Requires AVX512F.

func MaskzFnmsubSs 

func MaskzFnmsubSs(k x86.Mmask8, a x86.M128, b x86.M128, c x86.M128) (dst x86.M128)

MaskzFnmsubSs: Multiply the lower single-precision (32-bit) floating-point elements in 'a' and 'b', and subtract the lower element in 'c' from the negated intermediate result. Store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := -(a[31:0] * b[31:0]) - c[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VFNMSUB132SS, VFNMSUB213SS, VFNMSUB231SS'. Intrinsic: '_mm_maskz_fnmsub_ss'. Requires AVX512F.

func MaskzGetexpPd 

func MaskzGetexpPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzGetexpPd: Convert the exponent of each packed double-precision (64-bit) floating-point element in 'a' to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ConvertExpFP64(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPD'. Intrinsic: '_mm_maskz_getexp_pd'. Requires AVX512F.

func MaskzGetexpPs 

func MaskzGetexpPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzGetexpPs: Convert the exponent of each packed single-precision (32-bit) floating-point element in 'a' to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates 'floor(log2(x))' for each element. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ConvertExpFP32(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VGETEXPPS'. Intrinsic: '_mm_maskz_getexp_ps'. Requires AVX512F.

func MaskzGetexpRoundSd 

func MaskzGetexpRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzGetexpRoundSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := ConvertExpFP64(b[63:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_maskz_getexp_round_sd'. Requires AVX512F.

func MaskzGetexpRoundSs 

func MaskzGetexpRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzGetexpRoundSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := ConvertExpFP32(b[31:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_maskz_getexp_round_ss'. Requires AVX512F.

func MaskzGetexpSd 

func MaskzGetexpSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzGetexpSd: Convert the exponent of the lower double-precision (64-bit) floating-point element in 'b' to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

IF k[0]
	dst[63:0] := ConvertExpFP64(b[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VGETEXPSD'. Intrinsic: '_mm_maskz_getexp_sd'. Requires AVX512F.

func MaskzGetexpSs 

func MaskzGetexpSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzGetexpSs: Convert the exponent of the lower single-precision (32-bit) floating-point element in 'b' to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. This intrinsic essentially calculates 'floor(log2(x))' for the lower element. 

IF k[0]
	dst[31:0] := ConvertExpFP32(b[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VGETEXPSS'. Intrinsic: '_mm_maskz_getexp_ss'. Requires AVX512F.

func MaskzGetmantPd 

func MaskzGetmantPd(k x86.Mmask8, a x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

MaskzGetmantPd: Normalize the mantissas of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 1
			i := j*64
			IF k[j]
				dst[i+63:i] := GetNormalizedMantissa(a[i+63:i], sc, interv)
			ELSE
				dst[i+63:i] := 0
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPD'. Intrinsic: '_mm_maskz_getmant_pd'. Requires AVX512F.

func MaskzGetmantPs 

func MaskzGetmantPs(k x86.Mmask8, a x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

MaskzGetmantPs: Normalize the mantissas of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		FOR j := 0 to 3
			i := j*32
			IF k[j]
				dst[i+31:i] := GetNormalizedMantissa(a[i+31:i], sc, interv)
			ELSE
				dst[i+31:i] := 0
			FI
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VGETMANTPS'. Intrinsic: '_mm_maskz_getmant_ps'. Requires AVX512F.

func MaskzGetmantRoundSd 

func MaskzGetmantRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128d)

MaskzGetmantRoundSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_maskz_getmant_round_sd'. Requires AVX512F.

func MaskzGetmantRoundSs 

func MaskzGetmantRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM, rounding int) (dst x86.M128)

MaskzGetmantRoundSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1Rounding is done according to the 'rounding' parameter, which can be one of:
    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_maskz_getmant_round_ss'. Requires AVX512F.

func MaskzGetmantSd 

func MaskzGetmantSd(k x86.Mmask8, a x86.M128d, b x86.M128d, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128d)

MaskzGetmantSd: Normalize the mantissas of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		IF k[0]
			dst[63:0] := GetNormalizedMantissa(a[63:0], sc, interv)
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSD'. Intrinsic: '_mm_maskz_getmant_sd'. Requires AVX512F.

func MaskzGetmantSs 

func MaskzGetmantSs(k x86.Mmask8, a x86.M128, b x86.M128, interv MMMANTISSANORMENUM, sc MMMANTISSASIGNENUM) (dst x86.M128)

MaskzGetmantSs: Normalize the mantissas of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. This intrinsic essentially calculates '±(2^k)*|x.significand|', where 'k' depends on the interval range defined by 'interv' and the sign depends on 'sc' and the source sign. 

The mantissa is normalized to the interval specified by 'interv', which can

take the following values: 

    _MM_MANT_NORM_1_2     // interval [1, 2)
    _MM_MANT_NORM_p5_2    // interval [0.5, 2)
    _MM_MANT_NORM_p5_1    // interval [0.5, 1)
    _MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)The sign is determined by 'sc' which can take the following values:
    _MM_MANT_SIGN_src     // sign = sign(src)
    _MM_MANT_SIGN_zero    // sign = 0
    _MM_MANT_SIGN_nan     // dst = NaN if sign(src) = 1

		IF k[0]
			dst[31:0] := GetNormalizedMantissa(a[31:0], sc, interv)
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VGETMANTSS'. Intrinsic: '_mm_maskz_getmant_ss'. Requires AVX512F.

func MaskzLoadSd 

func MaskzLoadSd(k x86.Mmask8, mem_addr *float64) (dst x86.M128d)

MaskzLoadSd: Load a double-precision (64-bit) floating-point element from memory into the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and set the upper element of 'dst' to zero. 'mem_addr' must be aligned on a 16-byte boundary or a general-protection exception may be generated. 

IF k[0]
	dst[63:0] := MEM[mem_addr+63:mem_addr]
ELSE
	dst[63:0] := 0
FI
dst[MAX:64] := 0

Instruction: 'VMOVSD'. Intrinsic: '_mm_maskz_load_sd'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskzLoadSs 

func MaskzLoadSs(k x86.Mmask8, mem_addr *float32) (dst x86.M128)

MaskzLoadSs: Load a single-precision (32-bit) floating-point element from memory into the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and set the upper elements of 'dst' to zero. 'mem_addr' must be aligned on a 16-byte boundary or a general-protection exception may be generated. 

IF k[0]
	dst[31:0] := MEM[mem_addr+31:mem_addr]
ELSE
	dst[31:0] := 0
FI
dst[MAX:32] := 0

Instruction: 'VMOVSS'. Intrinsic: '_mm_maskz_load_ss'. Requires AVX512F.

FIXME: Will likely need to be reworked (has pointer parameter).

func MaskzMaxEpi32 

func MaskzMaxEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMaxEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXSD'. Intrinsic: '_mm_maskz_max_epi32'. Requires AVX512F.

func MaskzMaxEpi64 

func MaskzMaxEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm_maskz_max_epi64'. Requires AVX512F.

func MaskzMaxEpu32 

func MaskzMaxEpu32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMaxEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] > b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXUD'. Intrinsic: '_mm_maskz_max_epu32'. Requires AVX512F.

func MaskzMaxEpu64 

func MaskzMaxEpu64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] > b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm_maskz_max_epu64'. Requires AVX512F.

func MaskzMaxPd 

func MaskzMaxPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMaxPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := MAX(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMAXPD'. Intrinsic: '_mm_maskz_max_pd'. Requires AVX512F.

func MaskzMaxPs 

func MaskzMaxPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMaxPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed maximum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := MAX(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMAXPS'. Intrinsic: '_mm_maskz_max_ps'. Requires AVX512F.

func MaskzMaxRoundSd 

func MaskzMaxRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MaskzMaxRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[63:0] := MAX(a[63:0], b[63:0])
	ELSE
		dst[63:0] := 0
	FI
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMAXSD'. Intrinsic: '_mm_maskz_max_round_sd'. Requires AVX512F.

func MaskzMaxRoundSs 

func MaskzMaxRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, sae int) (dst x86.M128)

MaskzMaxRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[31:0] := MAX(a[31:0], b[31:0])
	ELSE
		dst[31:0] := 0
	FI
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMAXSS'. Intrinsic: '_mm_maskz_max_round_ss'. Requires AVX512F.

func MaskzMaxSd 

func MaskzMaxSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMaxSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := MAX(a[63:0], b[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMAXSD'. Intrinsic: '_mm_maskz_max_sd'. Requires AVX512F.

func MaskzMaxSs 

func MaskzMaxSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMaxSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[31:0] := MAX(a[31:0], b[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMAXSS'. Intrinsic: '_mm_maskz_max_ss'. Requires AVX512F.

func MaskzMinEpi32 

func MaskzMinEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMinEpi32: Compare packed 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINSD'. Intrinsic: '_mm_maskz_min_epi32'. Requires AVX512F.

func MaskzMinEpi64 

func MaskzMinEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm_maskz_min_epi64'. Requires AVX512F.

func MaskzMinEpu32 

func MaskzMinEpu32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMinEpu32: Compare packed unsigned 32-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF a[i+31:i] < b[i+31:i]
			dst[i+31:i] := a[i+31:i]
		ELSE
			dst[i+31:i] := b[i+31:i]
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINUD'. Intrinsic: '_mm_maskz_min_epu32'. Requires AVX512F.

func MaskzMinEpu64 

func MaskzMinEpu64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF a[i+63:i] < b[i+63:i]
			dst[i+63:i] := a[i+63:i]
		ELSE
			dst[i+63:i] := b[i+63:i]
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm_maskz_min_epu64'. Requires AVX512F.

func MaskzMinPd 

func MaskzMinPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMinPd: Compare packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := MIN(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMINPD'. Intrinsic: '_mm_maskz_min_pd'. Requires AVX512F.

func MaskzMinPs 

func MaskzMinPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMinPs: Compare packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store packed minimum values in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := MIN(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMINPS'. Intrinsic: '_mm_maskz_min_ps'. Requires AVX512F.

func MaskzMinRoundSd 

func MaskzMinRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MaskzMinRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[63:0] := MIN(a[63:0], b[63:0])
	ELSE
		dst[63:0] := 0
	FI
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMINSD'. Intrinsic: '_mm_maskz_min_round_sd'. Requires AVX512F.

func MaskzMinRoundSs 

func MaskzMinRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, sae int) (dst x86.M128)

MaskzMinRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	IF k[0]
		dst[31:0] := MIN(a[31:0], b[31:0])
	ELSE
		dst[31:0] := 0
	FI
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMINSS'. Intrinsic: '_mm_maskz_min_round_ss'. Requires AVX512F.

func MaskzMinSd 

func MaskzMinSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMinSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := MIN(a[63:0], b[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMINSD'. Intrinsic: '_mm_maskz_min_sd'. Requires AVX512F.

func MaskzMinSs 

func MaskzMinSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMinSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[31:0] := MIN(a[31:0], b[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMINSS'. Intrinsic: '_mm_maskz_min_ss'. Requires AVX512F.

func MaskzMovEpi32 

func MaskzMovEpi32(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzMovEpi32: Move packed 32-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDQA32'. Intrinsic: '_mm_maskz_mov_epi32'. Requires AVX512F.

func MaskzMovEpi64 

func MaskzMovEpi64(k x86.Mmask8, a x86.M128i) (dst x86.M128i)

MaskzMovEpi64: Move packed 64-bit integers from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDQA64'. Intrinsic: '_mm_maskz_mov_epi64'. Requires AVX512F.

func MaskzMovPd 

func MaskzMovPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzMovPd: Move packed double-precision (64-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVAPD'. Intrinsic: '_mm_maskz_mov_pd'. Requires AVX512F.

func MaskzMovPs 

func MaskzMovPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzMovPs: Move packed single-precision (32-bit) floating-point elements from 'a' into 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVAPS'. Intrinsic: '_mm_maskz_mov_ps'. Requires AVX512F.

func MaskzMoveSd 

func MaskzMoveSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMoveSd: Move the lower double-precision (64-bit) floating-point element from 'b' to the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := b[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMOVSD'. Intrinsic: '_mm_maskz_move_sd'. Requires AVX512F.

func MaskzMoveSs 

func MaskzMoveSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMoveSs: Move the lower single-precision (32-bit) floating-point element from 'b' to the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := b[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMOVSS'. Intrinsic: '_mm_maskz_move_ss'. Requires AVX512F.

func MaskzMovedupPd 

func MaskzMovedupPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzMovedupPd: Duplicate even-indexed double-precision (64-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[63:0] := a[63:0]
tmp[127:64] := a[63:0]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVDDUP'. Intrinsic: '_mm_maskz_movedup_pd'. Requires AVX512F.

func MaskzMovehdupPs 

func MaskzMovehdupPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzMovehdupPs: Duplicate odd-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[63:32]
tmp[63:32] := a[63:32]
tmp[95:64] := a[127:96]
tmp[127:96] := a[127:96]
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVSHDUP'. Intrinsic: '_mm_maskz_movehdup_ps'. Requires AVX512F.

func MaskzMoveldupPs 

func MaskzMoveldupPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzMoveldupPs: Duplicate even-indexed single-precision (32-bit) floating-point elements from 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp[31:0] := a[31:0]
tmp[63:32] := a[31:0]
tmp[95:64] := a[95:64]
tmp[127:96] := a[95:64]
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMOVSLDUP'. Intrinsic: '_mm_maskz_moveldup_ps'. Requires AVX512F.

func MaskzMulEpi32 

func MaskzMulEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMulEpi32: Multiply the low 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the signed 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULDQ'. Intrinsic: '_mm_maskz_mul_epi32'. Requires AVX512F.

func MaskzMulEpu32 

func MaskzMulEpu32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMulEpu32: Multiply the low unsigned 32-bit integers from each packed 64-bit element in 'a' and 'b', and store the unsigned 64-bit results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULUDQ'. Intrinsic: '_mm_maskz_mul_epu32'. Requires AVX512F.

func MaskzMulPd 

func MaskzMulPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMulPd: Multiply packed double-precision (64-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] * b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMULPD'. Intrinsic: '_mm_maskz_mul_pd'. Requires AVX512F.

func MaskzMulPs 

func MaskzMulPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMulPs: Multiply packed single-precision (32-bit) floating-point elements in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] * b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VMULPS'. Intrinsic: '_mm_maskz_mul_ps'. Requires AVX512F.

func MaskzMulRoundSd 

func MaskzMulRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzMulRoundSd: Multiply the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] * b[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VMULSD'. Intrinsic: '_mm_maskz_mul_round_sd'. Requires AVX512F.

func MaskzMulRoundSs 

func MaskzMulRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzMulRoundSs: Multiply the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] * b[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VMULSS'. Intrinsic: '_mm_maskz_mul_round_ss'. Requires AVX512F.

func MaskzMulSd 

func MaskzMulSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzMulSd: Multiply the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] * b[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VMULSD'. Intrinsic: '_mm_maskz_mul_sd'. Requires AVX512F.

func MaskzMulSs 

func MaskzMulSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzMulSs: Multiply the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] * b[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VMULSS'. Intrinsic: '_mm_maskz_mul_ss'. Requires AVX512F.

func MaskzMulloEpi32 

func MaskzMulloEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzMulloEpi32: Multiply the packed 32-bit integers in 'a' and 'b', producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		tmp[63:0] := a[i+31:i] * b[i+31:i]
		dst[i+31:i] := tmp[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMULLD'. Intrinsic: '_mm_maskz_mullo_epi32'. Requires AVX512F.

func MaskzOrEpi32 

func MaskzOrEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzOrEpi32: Compute the bitwise OR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] OR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPORD'. Intrinsic: '_mm_maskz_or_epi32'. Requires AVX512F.

func MaskzOrEpi64 

func MaskzOrEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzOrEpi64: Compute the bitwise OR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] OR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPORQ'. Intrinsic: '_mm_maskz_or_epi64'. Requires AVX512F.

func MaskzPermutePd 

func MaskzPermutePd(k x86.Mmask8, a x86.M128d, imm8 byte) (dst x86.M128d)

MaskzPermutePd: Shuffle double-precision (64-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (imm8[0] == 0) tmp_dst[63:0] := a[63:0]
IF (imm8[0] == 1) tmp_dst[63:0] := a[127:64]
IF (imm8[1] == 0) tmp_dst[127:64] := a[63:0]
IF (imm8[1] == 1) tmp_dst[127:64] := a[127:64]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm_maskz_permute_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzPermutePs 

func MaskzPermutePs(k x86.Mmask8, a x86.M128, imm8 byte) (dst x86.M128)

MaskzPermutePs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm_maskz_permute_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzPermutevarPd 

func MaskzPermutevarPd(k x86.Mmask8, a x86.M128d, b x86.M128i) (dst x86.M128d)

MaskzPermutevarPd: Shuffle double-precision (64-bit) floating-point elements in 'a' using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

IF (b[1] == 0) tmp_dst[63:0] := a[63:0]
IF (b[1] == 1) tmp_dst[63:0] := a[127:64]
IF (b[65] == 0) tmp_dst[127:64] := a[63:0]
IF (b[65] == 1) tmp_dst[127:64] := a[127:64]
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPD'. Intrinsic: '_mm_maskz_permutevar_pd'. Requires AVX512F.

func MaskzPermutevarPs 

func MaskzPermutevarPs(k x86.Mmask8, a x86.M128, b x86.M128i) (dst x86.M128)

MaskzPermutevarPs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], b[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], b[33:32])
tmp_dst[95:64] := SELECT4(a[127:0], b[65:64])
tmp_dst[127:96] := SELECT4(a[127:0], b[97:96])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMILPS'. Intrinsic: '_mm_maskz_permutevar_ps'. Requires AVX512F.

func MaskzPermutex2varEpi32 

func MaskzPermutex2varEpi32(k x86.Mmask8, a x86.M128i, idx x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzPermutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+2]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm_maskz_permutex2var_epi32'. Requires AVX512F.

func MaskzPermutex2varEpi64 

func MaskzPermutex2varEpi64(k x86.Mmask8, a x86.M128i, idx x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzPermutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+1]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm_maskz_permutex2var_epi64'. Requires AVX512F.

func MaskzPermutex2varPd 

func MaskzPermutex2varPd(k x86.Mmask8, a x86.M128d, idx x86.M128i, b x86.M128d) (dst x86.M128d)

MaskzPermutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	IF k[j]
		dst[i+63:i] := (idx[i+1]) ? b[off+63:off] : a[off+63:off]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm_maskz_permutex2var_pd'. Requires AVX512F.

func MaskzPermutex2varPs 

func MaskzPermutex2varPs(k x86.Mmask8, a x86.M128, idx x86.M128i, b x86.M128) (dst x86.M128)

MaskzPermutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	IF k[j]
		dst[i+31:i] := (idx[i+2]) ? b[off+31:off] : a[off+31:off]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm_maskz_permutex2var_ps'. Requires AVX512F.

func MaskzRcp14Pd 

func MaskzRcp14Pd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzRcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm_maskz_rcp14_pd'. Requires AVX512F.

func MaskzRcp14Ps 

func MaskzRcp14Ps(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzRcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm_maskz_rcp14_ps'. Requires AVX512F.

func MaskzRcp14Sd 

func MaskzRcp14Sd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzRcp14Sd: Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[63:0] := APPROXIMATE(1.0/b[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRCP14SD'. Intrinsic: '_mm_maskz_rcp14_sd'. Requires AVX512F.

func MaskzRcp14Ss 

func MaskzRcp14Ss(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzRcp14Ss: Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[31:0] := APPROXIMATE(1.0/b[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRCP14SS'. Intrinsic: '_mm_maskz_rcp14_ss'. Requires AVX512F.

func MaskzRolEpi32 

func MaskzRolEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzRolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm_maskz_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRolEpi64 

func MaskzRolEpi64(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzRolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm_maskz_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRolvEpi32 

func MaskzRolvEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzRolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm_maskz_rolv_epi32'. Requires AVX512F.

func MaskzRolvEpi64 

func MaskzRolvEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzRolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm_maskz_rolv_epi64'. Requires AVX512F.

func MaskzRorEpi32 

func MaskzRorEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzRorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm_maskz_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRorEpi64 

func MaskzRorEpi64(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzRorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm_maskz_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRorvEpi32 

func MaskzRorvEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzRorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm_maskz_rorv_epi32'. Requires AVX512F.

func MaskzRorvEpi64 

func MaskzRorvEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzRorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm_maskz_rorv_epi64'. Requires AVX512F.

func MaskzRoundscalePd 

func MaskzRoundscalePd(k x86.Mmask8, a x86.M128d, imm8 byte) (dst x86.M128d)

MaskzRoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm_maskz_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRoundscalePs 

func MaskzRoundscalePs(k x86.Mmask8, a x86.M128, imm8 byte) (dst x86.M128)

MaskzRoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm_maskz_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRoundscaleRoundSd 

func MaskzRoundscaleRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte, rounding int) (dst x86.M128d)

MaskzRoundscaleRoundSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		IF k[0]
			dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_maskz_roundscale_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRoundscaleRoundSs 

func MaskzRoundscaleRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte, rounding int) (dst x86.M128)

MaskzRoundscaleRoundSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		IF k[0]
			dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_maskz_roundscale_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRoundscaleSd 

func MaskzRoundscaleSd(k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.M128d)

MaskzRoundscaleSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

IF k[0]
	dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_maskz_roundscale_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRoundscaleSs 

func MaskzRoundscaleSs(k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.M128)

MaskzRoundscaleSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

IF k[0]
	dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_maskz_roundscale_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzRsqrt14Pd 

func MaskzRsqrt14Pd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzRsqrt14Pd: Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := APPROXIMATE(1.0 / SQRT(a[i+63:i]))
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRSQRT14PD'. Intrinsic: '_mm_maskz_rsqrt14_pd'. Requires AVX512F.

func MaskzRsqrt14Ps 

func MaskzRsqrt14Ps(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzRsqrt14Ps: Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := APPROXIMATE(1.0 / SQRT(a[i+31:i]))
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRSQRT14PS'. Intrinsic: '_mm_maskz_rsqrt14_ps'. Requires AVX512F.

func MaskzRsqrt14Sd 

func MaskzRsqrt14Sd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzRsqrt14Sd: Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[63:0] := APPROXIMATE(1.0 / SQRT(b[63:0]))
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SD'. Intrinsic: '_mm_maskz_rsqrt14_sd'. Requires AVX512F.

func MaskzRsqrt14Ss 

func MaskzRsqrt14Ss(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzRsqrt14Ss: Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

IF k[0]
	dst[31:0] := APPROXIMATE(1.0 / SQRT(b[31:0]))
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SS'. Intrinsic: '_mm_maskz_rsqrt14_ss'. Requires AVX512F.

func MaskzScalefPd 

func MaskzScalefPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm_maskz_scalef_pd'. Requires AVX512F.

func MaskzScalefPs 

func MaskzScalefPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm_maskz_scalef_ps'. Requires AVX512F.

func MaskzScalefRoundSd 

func MaskzScalefRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzScalefRoundSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		IF k[0]
			dst[63:0] := SCALE(a[63:0], b[63:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_maskz_scalef_round_sd'. Requires AVX512F.

func MaskzScalefRoundSs 

func MaskzScalefRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzScalefRoundSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[63:0]
		}

		IF k[0]
			dst[31:0] := SCALE(a[31:0], b[31:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_maskz_scalef_round_ss'. Requires AVX512F.

func MaskzScalefSd 

func MaskzScalefSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzScalefSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

IF k[0]
	dst[63:0] := SCALE(a[63:0], b[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_maskz_scalef_sd'. Requires AVX512F.

func MaskzScalefSs 

func MaskzScalefSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzScalefSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[63:0]
}

IF k[0]
	dst[31:0] := SCALE(a[31:0], b[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_maskz_scalef_ss'. Requires AVX512F.

func MaskzSet1Epi32 

func MaskzSet1Epi32(k x86.Mmask8, a int) (dst x86.M128i)

MaskzSet1Epi32: Broadcast 32-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[31:0]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTD'. Intrinsic: '_mm_maskz_set1_epi32'. Requires AVX512F.

func MaskzSet1Epi64 

func MaskzSet1Epi64(k x86.Mmask8, a int64) (dst x86.M128i)

MaskzSet1Epi64: Broadcast 64-bit integer 'a' to all elements of 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[63:0]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPBROADCASTQ'. Intrinsic: '_mm_maskz_set1_epi64'. Requires AVX512F.

func MaskzShuffleEpi32 

func MaskzShuffleEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzShuffleEpi32: Shuffle 32-bit integers in 'a' using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(a[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(a[127:0], imm8[7:6])
FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSHUFD'. Intrinsic: '_mm_maskz_shuffle_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzShufflePd 

func MaskzShufflePd(k x86.Mmask8, a x86.M128d, b x86.M128d, imm8 byte) (dst x86.M128d)

MaskzShufflePd: Shuffle double-precision (64-bit) floating-point elements using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

tmp_dst[63:0] := (imm8[0] == 0) ? a[63:0] : a[127:64]
tmp_dst[127:64] := (imm8[1] == 0) ? b[63:0] : b[127:64]

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSHUFPD'. Intrinsic: '_mm_maskz_shuffle_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzShufflePs 

func MaskzShufflePs(k x86.Mmask8, a x86.M128, b x86.M128, imm8 byte) (dst x86.M128)

MaskzShufflePs: Shuffle single-precision (32-bit) floating-point elements in 'a' using the control in 'imm8', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

SELECT4(src, control){
	CASE(control[1:0])
	0:	tmp[31:0] := src[31:0]
	1:	tmp[31:0] := src[63:32]
	2:	tmp[31:0] := src[95:64]
	3:	tmp[31:0] := src[127:96]
	ESAC
	RETURN tmp[31:0]
}

tmp_dst[31:0] := SELECT4(a[127:0], imm8[1:0])
tmp_dst[63:32] := SELECT4(a[127:0], imm8[3:2])
tmp_dst[95:64] := SELECT4(b[127:0], imm8[5:4])
tmp_dst[127:96] := SELECT4(b[127:0], imm8[7:6])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSHUFPS'. Intrinsic: '_mm_maskz_shuffle_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSllEpi32 

func MaskzSllEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSllEpi32: Shift packed 32-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm_maskz_sll_epi32'. Requires AVX512F.

func MaskzSllEpi64 

func MaskzSllEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSllEpi64: Shift packed 64-bit integers in 'a' left by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm_maskz_sll_epi64'. Requires AVX512F.

func MaskzSlliEpi32 

func MaskzSlliEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSlliEpi32: Shift packed 32-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] << imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLD'. Intrinsic: '_mm_maskz_slli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSlliEpi64 

func MaskzSlliEpi64(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSlliEpi64: Shift packed 64-bit integers in 'a' left by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLQ'. Intrinsic: '_mm_maskz_slli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSllvEpi32 

func MaskzSllvEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSllvEpi32: Shift packed 32-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] << count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLVD'. Intrinsic: '_mm_maskz_sllv_epi32'. Requires AVX512F.

func MaskzSllvEpi64 

func MaskzSllvEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSllvEpi64: Shift packed 64-bit integers in 'a' left by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] << count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSLLVQ'. Intrinsic: '_mm_maskz_sllv_epi64'. Requires AVX512F.

func MaskzSqrtPd 

func MaskzSqrtPd(k x86.Mmask8, a x86.M128d) (dst x86.M128d)

MaskzSqrtPd: Compute the square root of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SQRT(a[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSQRTPD'. Intrinsic: '_mm_maskz_sqrt_pd'. Requires AVX512F.

func MaskzSqrtPs 

func MaskzSqrtPs(k x86.Mmask8, a x86.M128) (dst x86.M128)

MaskzSqrtPs: Compute the square root of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SQRT(a[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSQRTPS'. Intrinsic: '_mm_maskz_sqrt_ps'. Requires AVX512F.

func MaskzSqrtRoundSd 

func MaskzSqrtRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzSqrtRoundSd: Compute the square root of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := SQRT(a[63:0])
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSQRTSD'. Intrinsic: '_mm_maskz_sqrt_round_sd'. Requires AVX512F.

func MaskzSqrtRoundSs 

func MaskzSqrtRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzSqrtRoundSs: Compute the square root of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := SQRT(a[31:0])
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSQRTSS'. Intrinsic: '_mm_maskz_sqrt_round_ss'. Requires AVX512F.

func MaskzSqrtSd 

func MaskzSqrtSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzSqrtSd: Compute the square root of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'b' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := SQRT(a[63:0])
ELSE
	dst[63:0] := 0
FI
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VSQRTSD'. Intrinsic: '_mm_maskz_sqrt_sd'. Requires AVX512F.

func MaskzSqrtSs 

func MaskzSqrtSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzSqrtSs: Compute the square root of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := SQRT(a[31:0])
ELSE
	dst[31:0] := 0
FI
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VSQRTSS'. Intrinsic: '_mm_maskz_sqrt_ss'. Requires AVX512F.

func MaskzSraEpi32 

func MaskzSraEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSraEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm_maskz_sra_epi32'. Requires AVX512F.

func MaskzSraEpi64 

func MaskzSraEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_maskz_sra_epi64'. Requires AVX512F.

func MaskzSraiEpi32 

func MaskzSraiEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSraiEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := SignBit
		ELSE
			dst[i+31:i] := SignExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAD'. Intrinsic: '_mm_maskz_srai_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSraiEpi64 

func MaskzSraiEpi64(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := SignBit
		ELSE
			dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_maskz_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSravEpi32 

func MaskzSravEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSravEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := SignExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAVD'. Intrinsic: '_mm_maskz_srav_epi32'. Requires AVX512F.

func MaskzSravEpi64 

func MaskzSravEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm_maskz_srav_epi64'. Requires AVX512F.

func MaskzSrlEpi32 

func MaskzSrlEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSrlEpi32: Shift packed 32-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF count[63:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[63:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm_maskz_srl_epi32'. Requires AVX512F.

func MaskzSrlEpi64 

func MaskzSrlEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSrlEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF count[63:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[63:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm_maskz_srl_epi64'. Requires AVX512F.

func MaskzSrliEpi32 

func MaskzSrliEpi32(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSrliEpi32: Shift packed 32-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		IF imm8[7:0] > 31
			dst[i+31:i] := 0
		ELSE
			dst[i+31:i] := ZeroExtend(a[i+31:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLD'. Intrinsic: '_mm_maskz_srli_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSrliEpi64 

func MaskzSrliEpi64(k x86.Mmask8, a x86.M128i, imm8 byte) (dst x86.M128i)

MaskzSrliEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		IF imm8[7:0] > 63
			dst[i+63:i] := 0
		ELSE
			dst[i+63:i] := ZeroExtend(a[i+63:i] >> imm8[7:0])
		FI
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLQ'. Intrinsic: '_mm_maskz_srli_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzSrlvEpi32 

func MaskzSrlvEpi32(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSrlvEpi32: Shift packed 32-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := ZeroExtend(a[i+31:i] >> count[i+31:i])
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLVD'. Intrinsic: '_mm_maskz_srlv_epi32'. Requires AVX512F.

func MaskzSrlvEpi64 

func MaskzSrlvEpi64(k x86.Mmask8, a x86.M128i, count x86.M128i) (dst x86.M128i)

MaskzSrlvEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in zeros, and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := ZeroExtend(a[i+63:i] >> count[i+63:i])
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRLVQ'. Intrinsic: '_mm_maskz_srlv_epi64'. Requires AVX512F.

func MaskzSubEpi32 

func MaskzSubEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzSubEpi32: Subtract packed 32-bit integers in 'b' from packed 32-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSUBD'. Intrinsic: '_mm_maskz_sub_epi32'. Requires AVX512F.

func MaskzSubEpi64 

func MaskzSubEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzSubEpi64: Subtract packed 64-bit integers in 'b' from packed 64-bit integers in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSUBQ'. Intrinsic: '_mm_maskz_sub_epi64'. Requires AVX512F.

func MaskzSubPd 

func MaskzSubPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzSubPd: Subtract packed double-precision (64-bit) floating-point elements in 'b' from packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] - b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSUBPD'. Intrinsic: '_mm_maskz_sub_pd'. Requires AVX512F.

func MaskzSubPs 

func MaskzSubPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzSubPs: Subtract packed single-precision (32-bit) floating-point elements in 'b' from packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] - b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSUBPS'. Intrinsic: '_mm_maskz_sub_ps'. Requires AVX512F.

func MaskzSubRoundSd 

func MaskzSubRoundSd(k x86.Mmask8, a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MaskzSubRoundSd: Subtract the lower double-precision (64-bit) floating-point element in 'b' from the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[63:0] := a[63:0] - b[63:0]
		ELSE
			dst[63:0] := 0
		FI
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VSUBSD'. Intrinsic: '_mm_maskz_sub_round_sd'. Requires AVX512F.

func MaskzSubRoundSs 

func MaskzSubRoundSs(k x86.Mmask8, a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MaskzSubRoundSs: Subtract the lower single-precision (32-bit) floating-point element in 'b' from the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		IF k[0]
			dst[31:0] := a[31:0] - b[31:0]
		ELSE
			dst[31:0] := 0
		FI
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VSUBSS'. Intrinsic: '_mm_maskz_sub_round_ss'. Requires AVX512F.

func MaskzSubSd 

func MaskzSubSd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzSubSd: Subtract the lower double-precision (64-bit) floating-point element in 'b' from the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper element from 'a' to the upper element of 'dst'. 

IF k[0]
	dst[63:0] := a[63:0] - b[63:0]
ELSE
	dst[63:0] := 0
FI
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VSUBSD'. Intrinsic: '_mm_maskz_sub_sd'. Requires AVX512F.

func MaskzSubSs 

func MaskzSubSs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzSubSs: Subtract the lower single-precision (32-bit) floating-point element in 'b' from the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst' using zeromask 'k' (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

IF k[0]
	dst[31:0] := a[31:0] - b[31:0]
ELSE
	dst[31:0] := 0
FI
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VSUBSS'. Intrinsic: '_mm_maskz_sub_ss'. Requires AVX512F.

func MaskzTernarylogicEpi32 

func MaskzTernarylogicEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i, c x86.M128i, imm8 byte) (dst x86.M128i)

MaskzTernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		FOR h := 0 to 31
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm_maskz_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzTernarylogicEpi64 

func MaskzTernarylogicEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i, c x86.M128i, imm8 byte) (dst x86.M128i)

MaskzTernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst' using zeromask 'k' at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		FOR h := 0 to 63
			index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
			dst[i+h] := imm8[index[2:0]]
		ENDFOR
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm_maskz_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func MaskzUnpackhiEpi32 

func MaskzUnpackhiEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzUnpackhiEpi32: Unpack and interleave 32-bit integers from the high half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKHDQ'. Intrinsic: '_mm_maskz_unpackhi_epi32'. Requires AVX512F.

func MaskzUnpackhiEpi64 

func MaskzUnpackhiEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzUnpackhiEpi64: Unpack and interleave 64-bit integers from the high half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKHQDQ'. Intrinsic: '_mm_maskz_unpackhi_epi64'. Requires AVX512F.

func MaskzUnpackhiPd 

func MaskzUnpackhiPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzUnpackhiPd: Unpack and interleave double-precision (64-bit) floating-point elements from the high half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[127:64]
	dst[127:64] := src2[127:64]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKHPD'. Intrinsic: '_mm_maskz_unpackhi_pd'. Requires AVX512F.

func MaskzUnpackhiPs 

func MaskzUnpackhiPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzUnpackhiPs: Unpack and interleave single-precision (32-bit) floating-point elements from the high half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_HIGH_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[95:64]
	dst[63:32] := src2[95:64]
	dst[95:64] := src1[127:96]
	dst[127:96] := src2[127:96]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_HIGH_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKHPS'. Intrinsic: '_mm_maskz_unpackhi_ps'. Requires AVX512F.

func MaskzUnpackloEpi32 

func MaskzUnpackloEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzUnpackloEpi32: Unpack and interleave 32-bit integers from the low half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKLDQ'. Intrinsic: '_mm_maskz_unpacklo_epi32'. Requires AVX512F.

func MaskzUnpackloEpi64 

func MaskzUnpackloEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzUnpackloEpi64: Unpack and interleave 64-bit integers from the low half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPUNPCKLQDQ'. Intrinsic: '_mm_maskz_unpacklo_epi64'. Requires AVX512F.

func MaskzUnpackloPd 

func MaskzUnpackloPd(k x86.Mmask8, a x86.M128d, b x86.M128d) (dst x86.M128d)

MaskzUnpackloPd: Unpack and interleave double-precision (64-bit) floating-point elements from the low half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_QWORDS(src1[127:0], src2[127:0]){
	dst[63:0] := src1[63:0]
	dst[127:64] := src2[63:0]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_QWORDS(a[127:0], b[127:0])

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := tmp_dst[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKLPD'. Intrinsic: '_mm_maskz_unpacklo_pd'. Requires AVX512F.

func MaskzUnpackloPs 

func MaskzUnpackloPs(k x86.Mmask8, a x86.M128, b x86.M128) (dst x86.M128)

MaskzUnpackloPs: Unpack and interleave single-precision (32-bit) floating-point elements from the low half of 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

INTERLEAVE_DWORDS(src1[127:0], src2[127:0]){
	dst[31:0] := src1[31:0]
	dst[63:32] := src2[31:0]
	dst[95:64] := src1[63:32]
	dst[127:96] := src2[63:32]
	RETURN dst[127:0]
}

tmp_dst[127:0] := INTERLEAVE_DWORDS(a[127:0], b[127:0])

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := tmp_dst[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VUNPCKLPS'. Intrinsic: '_mm_maskz_unpacklo_ps'. Requires AVX512F.

func MaskzXorEpi32 

func MaskzXorEpi32(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzXorEpi32: Compute the bitwise XOR of packed 32-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 3
	i := j*32
	IF k[j]
		dst[i+31:i] := a[i+31:i] XOR b[i+31:i]
	ELSE
		dst[i+31:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPXORD'. Intrinsic: '_mm_maskz_xor_epi32'. Requires AVX512F.

func MaskzXorEpi64 

func MaskzXorEpi64(k x86.Mmask8, a x86.M128i, b x86.M128i) (dst x86.M128i)

MaskzXorEpi64: Compute the bitwise XOR of packed 64-bit integers in 'a' and 'b', and store the results in 'dst' using zeromask 'k' (elements are zeroed out when the corresponding mask bit is not set). 

FOR j := 0 to 1
	i := j*64
	IF k[j]
		dst[i+63:i] := a[i+63:i] XOR b[i+63:i]
	ELSE
		dst[i+63:i] := 0
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPXORQ'. Intrinsic: '_mm_maskz_xor_epi64'. Requires AVX512F.

func MaxEpi64 

func MaxEpi64(a x86.M128i, b x86.M128i) (dst x86.M128i)

MaxEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXSQ'. Intrinsic: '_mm_max_epi64'. Requires AVX512F.

func MaxEpu64 

func MaxEpu64(a x86.M128i, b x86.M128i) (dst x86.M128i)

MaxEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed maximum values in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF a[i+63:i] > b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMAXUQ'. Intrinsic: '_mm_max_epu64'. Requires AVX512F.

func MaxRoundSd 

func MaxRoundSd(a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MaxRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	dst[63:0] := MAX(a[63:0], b[63:0])
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMAXSD'. Intrinsic: '_mm_max_round_sd'. Requires AVX512F.

func MaxRoundSs 

func MaxRoundSs(a x86.M128, b x86.M128, sae int) (dst x86.M128)

MaxRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the maximum value in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	dst[31:0] := MAX(a[31:0], b[31:0])
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMAXSS'. Intrinsic: '_mm_max_round_ss'. Requires AVX512F.

func MinEpi64 

func MinEpi64(a x86.M128i, b x86.M128i) (dst x86.M128i)

MinEpi64: Compare packed 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINSQ'. Intrinsic: '_mm_min_epi64'. Requires AVX512F.

func MinEpu64 

func MinEpu64(a x86.M128i, b x86.M128i) (dst x86.M128i)

MinEpu64: Compare packed unsigned 64-bit integers in 'a' and 'b', and store packed minimum values in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF a[i+63:i] < b[i+63:i]
		dst[i+63:i] := a[i+63:i]
	ELSE
		dst[i+63:i] := b[i+63:i]
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPMINUQ'. Intrinsic: '_mm_min_epu64'. Requires AVX512F.

func MinRoundSd 

func MinRoundSd(a x86.M128d, b x86.M128d, sae int) (dst x86.M128d)

MinRoundSd: Compare the lower double-precision (64-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst' , and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	dst[63:0] := MIN(a[63:0], b[63:0])
	dst[127:64] := a[127:64]
	dst[MAX:128] := 0

Instruction: 'VMINSD'. Intrinsic: '_mm_min_round_sd'. Requires AVX512F.

func MinRoundSs 

func MinRoundSs(a x86.M128, b x86.M128, sae int) (dst x86.M128)

MinRoundSs: Compare the lower single-precision (32-bit) floating-point elements in 'a' and 'b', store the minimum value in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Pass __MM_FROUND_NO_EXC to 'sae' to suppress all exceptions.

	dst[31:0] := MIN(a[31:0], b[31:0])
	dst[127:32] := a[127:32]
	dst[MAX:128] := 0

Instruction: 'VMINSS'. Intrinsic: '_mm_min_round_ss'. Requires AVX512F.

func MulRoundSd 

func MulRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

MulRoundSd: Multiply the lower double-precision (64-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := a[63:0] * b[63:0]
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VMULSD'. Intrinsic: '_mm_mul_round_sd'. Requires AVX512F.

func MulRoundSs 

func MulRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

MulRoundSs: Multiply the lower single-precision (32-bit) floating-point element in 'a' and 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := a[31:0] * b[31:0]
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VMULSS'. Intrinsic: '_mm_mul_round_ss'. Requires AVX512F.

func Permutex2varEpi32 

func Permutex2varEpi32(a x86.M128i, idx x86.M128i, b x86.M128i) (dst x86.M128i)

Permutex2varEpi32: Shuffle 32-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*32
	off := idx[i+2:i]*32
	dst[i+31:i] := idx[i+3] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2D, VPERMT2D'. Intrinsic: '_mm_permutex2var_epi32'. Requires AVX512F.

func Permutex2varEpi64 

func Permutex2varEpi64(a x86.M128i, idx x86.M128i, b x86.M128i) (dst x86.M128i)

Permutex2varEpi64: Shuffle 64-bit integers in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2Q, VPERMT2Q'. Intrinsic: '_mm_permutex2var_epi64'. Requires AVX512F.

func Permutex2varPd 

func Permutex2varPd(a x86.M128d, idx x86.M128i, b x86.M128d) (dst x86.M128d)

Permutex2varPd: Shuffle double-precision (64-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	off := idx[i]*64
	dst[i+63:i] := idx[i+1] ? b[off+63:off] : a[off+63:off]
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PD, VPERMT2PD'. Intrinsic: '_mm_permutex2var_pd'. Requires AVX512F.

func Permutex2varPs 

func Permutex2varPs(a x86.M128, idx x86.M128i, b x86.M128) (dst x86.M128)

Permutex2varPs: Shuffle single-precision (32-bit) floating-point elements in 'a' and 'b' using the corresponding selector and index in 'idx', and store the results in 'dst'. 

FOR j := 0 to 3
	i := j*32
	off := idx[i+1:i]*32
	dst[i+31:i] := idx[i+2] ? b[off+31:off] : a[off+31:off]
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPERMI2PS, VPERMT2PS'. Intrinsic: '_mm_permutex2var_ps'. Requires AVX512F.

func Rcp14Pd 

func Rcp14Pd(a x86.M128d) (dst x86.M128d)

Rcp14Pd: Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := APPROXIMATE(1.0/a[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PD'. Intrinsic: '_mm_rcp14_pd'. Requires AVX512F.

func Rcp14Ps 

func Rcp14Ps(a x86.M128) (dst x86.M128)

Rcp14Ps: Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in 'a', and store the results in 'dst'. The maximum relative error for this approximation is less than 2^-14. 

FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := APPROXIMATE(1.0/a[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRCP14PS'. Intrinsic: '_mm_rcp14_ps'. Requires AVX512F.

func Rcp14Sd 

func Rcp14Sd(a x86.M128d, b x86.M128d) (dst x86.M128d)

Rcp14Sd: Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

dst[63:0] := APPROXIMATE(1.0/b[63:0])
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRCP14SD'. Intrinsic: '_mm_rcp14_sd'. Requires AVX512F.

func Rcp14Ss 

func Rcp14Ss(a x86.M128, b x86.M128) (dst x86.M128)

Rcp14Ss: Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

dst[31:0] := APPROXIMATE(1.0/b[31:0])
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRCP14SS'. Intrinsic: '_mm_rcp14_ss'. Requires AVX512F.

func RolEpi32 

func RolEpi32(a x86.M128i, imm8 byte) (dst x86.M128i)

RolEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLD'. Intrinsic: '_mm_rol_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RolEpi64 

func RolEpi64(a x86.M128i, imm8 byte) (dst x86.M128i)

RolEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in 'imm8', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLQ'. Intrinsic: '_mm_rol_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RolvEpi32 

func RolvEpi32(a x86.M128i, b x86.M128i) (dst x86.M128i)

RolvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src << count) OR (src >> (32 - count))
}
FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := LEFT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVD'. Intrinsic: '_mm_rolv_epi32'. Requires AVX512F.

func RolvEpi64 

func RolvEpi64(a x86.M128i, b x86.M128i) (dst x86.M128i)

RolvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the left by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

LEFT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src << count) OR (src >> (64 - count))
}
FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := LEFT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPROLVQ'. Intrinsic: '_mm_rolv_epi64'. Requires AVX512F.

func RorEpi32 

func RorEpi32(a x86.M128i, imm8 byte) (dst x86.M128i)

RorEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORD'. Intrinsic: '_mm_ror_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RorEpi64 

func RorEpi64(a x86.M128i, imm8 byte) (dst x86.M128i)

RorEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in 'imm8', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORQ'. Intrinsic: '_mm_ror_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RorvEpi32 

func RorvEpi32(a x86.M128i, b x86.M128i) (dst x86.M128i)

RorvEpi32: Rotate the bits in each packed 32-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_DWORDS(src, count_src){
	count := count_src modulo 32
	RETURN (src >>count) OR (src << (32 - count))
}
FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := RIGHT_ROTATE_DWORDS(a[i+31:i], b[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVD'. Intrinsic: '_mm_rorv_epi32'. Requires AVX512F.

func RorvEpi64 

func RorvEpi64(a x86.M128i, b x86.M128i) (dst x86.M128i)

RorvEpi64: Rotate the bits in each packed 64-bit integer in 'a' to the right by the number of bits specified in the corresponding element of 'b', and store the results in 'dst'. 

RIGHT_ROTATE_QWORDS(src, count_src){
	count := count_src modulo 64
	RETURN (src >> count) OR (src << (64 - count))
}
FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := RIGHT_ROTATE_QWORDS(a[i+63:i], b[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPRORVQ'. Intrinsic: '_mm_rorv_epi64'. Requires AVX512F.

func RoundscalePd 

func RoundscalePd(a x86.M128d, imm8 byte) (dst x86.M128d)

RoundscalePd: Round packed double-precision (64-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := RoundTo_IntegerPD(a[i+63:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPD'. Intrinsic: '_mm_roundscale_pd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RoundscalePs 

func RoundscalePs(a x86.M128, imm8 byte) (dst x86.M128)

RoundscalePs: Round packed single-precision (32-bit) floating-point elements in 'a' to the number of fraction bits specified by 'imm8', and store the results in 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := RoundTo_IntegerPS(a[i+31:i], imm8[7:0])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VRNDSCALEPS'. Intrinsic: '_mm_roundscale_ps'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RoundscaleRoundSd 

func RoundscaleRoundSd(a x86.M128d, b x86.M128d, imm8 byte, rounding int) (dst x86.M128d)

RoundscaleRoundSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPD(src[63:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
			1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
			2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
			3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
			ESAC

			dst[63:0] := 2^-M * tmp[63:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[63:0] != dst[63:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[63:0]
		}

		dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_roundscale_round_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RoundscaleRoundSs 

func RoundscaleRoundSs(a x86.M128, b x86.M128, imm8 byte, rounding int) (dst x86.M128)

RoundscaleRoundSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		RoundTo_IntegerPS(src[31:0], imm8[7:0]){
			IF(imm8[2] == 1)
				rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
			ELSE
				rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
			FI

			M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

			CASE(rounding_direction)
			0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
			1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
			2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
			3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
			ESAC

			dst[31:0] := 2^-M * tmp[31:0] // scale back down

			IF imm8[3] == 0 //check SPE
				IF src[31:0] != dst[31:0] //check if precision has been lost
					set_precision() //set #PE
				FI
			FI
			RETURN dst[31:0]
		}

		dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_roundscale_round_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RoundscaleSd 

func RoundscaleSd(a x86.M128d, b x86.M128d, imm8 byte) (dst x86.M128d)

RoundscaleSd: Round the lower double-precision (64-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. 

RoundTo_IntegerPD(src[63:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[63:0] := round_to_nearest_even_integer(2^M * src[63:0])
	1: tmp[63:0] := round_to_equal_or_smaller_integer(2^M * src[63:0])
	2: tmp[63:0] := round_to_equal_or_larger_integer(2^M * src[63:0])
	3: tmp[63:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[63:0])
	ESAC

	dst[63:0] := 2^-M * tmp[63:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[63:0] != dst[63:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[63:0]
}

dst[63:0] := RoundTo_IntegerPD(a[63:0], imm8[7:0])
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESD'. Intrinsic: '_mm_roundscale_sd'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func RoundscaleSs 

func RoundscaleSs(a x86.M128, b x86.M128, imm8 byte) (dst x86.M128)

RoundscaleSs: Round the lower single-precision (32-bit) floating-point element in 'a' to the number of fraction bits specified by 'imm8', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

RoundTo_IntegerPS(src[31:0], imm8[7:0]){
	IF(imm8[2] == 1)
		rounding_direction := MXCSR.RC //Use the rounding mode specified by MXCSR.RC
	ELSE
		rounding_direction := imm8[1:0] //Use the rounding mode specified by imm8[1:0]
	FI

	M := imm8[7:4] // The scaling factor (number of fraction bits to round to)

	CASE(rounding_direction)
	0: tmp[31:0] := round_to_nearest_even_integer(2^M * src[31:0])
	1: tmp[31:0] := round_to_equal_or_smaller_integer(2^M * src[31:0])
	2: tmp[31:0] := round_to_equal_or_larger_integer(2^M * src[31:0])
	3: tmp[31:0] := round_to_nearest_smallest_magnitude_integer(2^M * src[31:0])
	ESAC

	dst[31:0] := 2^-M * tmp[31:0] // scale back down

	IF imm8[3] == 0 //check SPE
		IF src[31:0] != dst[31:0] //check if precision has been lost
			set_precision() //set #PE
		FI
	FI
	RETURN dst[31:0]
}

dst[31:0] := RoundTo_IntegerPS(a[31:0], imm8[7:0])
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VRNDSCALESS'. Intrinsic: '_mm_roundscale_ss'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func Rsqrt14Sd 

func Rsqrt14Sd(a x86.M128d, b x86.M128d) (dst x86.M128d)

Rsqrt14Sd: Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

dst[63:0] := APPROXIMATE(1.0 / SQRT(b[63:0]))
dst[127:64] := a[127:64]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SD'. Intrinsic: '_mm_rsqrt14_sd'. Requires AVX512F.

func Rsqrt14Ss 

func Rsqrt14Ss(a x86.M128, b x86.M128) (dst x86.M128)

Rsqrt14Ss: Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. The maximum relative error for this approximation is less than 2^-14. 

dst[31:0] := APPROXIMATE(1.0 / SQRT(b[31:0]))
dst[127:32] := a[127:32]
dst[MAX:128] := 0

Instruction: 'VRSQRT14SS'. Intrinsic: '_mm_rsqrt14_ss'. Requires AVX512F.

func ScalefPd 

func ScalefPd(a x86.M128d, b x86.M128d) (dst x86.M128d)

ScalefPd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := SCALE(a[i+63:0], b[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPD'. Intrinsic: '_mm_scalef_pd'. Requires AVX512F.

func ScalefPs 

func ScalefPs(a x86.M128, b x86.M128) (dst x86.M128)

ScalefPs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', and store the results in 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[31:0]
}

FOR j := 0 to 3
	i := j*32
	dst[i+31:i] := SCALE(a[i+31:0], b[i+31:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VSCALEFPS'. Intrinsic: '_mm_scalef_ps'. Requires AVX512F.

func ScalefRoundSd 

func ScalefRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

ScalefRoundSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
			RETURN dst[63:0]
		}

		dst[63:0] := SCALE(a[63:0], b[63:0])
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_scalef_round_sd'. Requires AVX512F.

func ScalefRoundSs 

func ScalefRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

ScalefRoundSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		SCALE(src1, src2){
			IF (src2 == NaN)
				IF (src2 == SNaN)
					RETURN QNAN(src2)
				FI
			ELSE IF (src1 == NaN)
				IF (src1 == SNaN)
					RETURN QNAN(src1)
				FI
				IF (src2 != INF)
					RETURN QNAN(src1)
				FI
			ELSE
				tmp_src2 := src2
				tmp_src1 := src1
				IF (src2 is denormal AND MXCSR.DAZ)
					tmp_src2 := 0
				FI
				IF (src1 is denormal AND MXCSR.DAZ)
					tmp_src1 := 0
				FI
			FI
			dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
			RETURN dst[63:0]
		}

		dst[31:0] := SCALE(a[31:0], b[31:0])
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_scalef_round_ss'. Requires AVX512F.

func ScalefSd 

func ScalefSd(a x86.M128d, b x86.M128d) (dst x86.M128d)

ScalefSd: Scale the packed double-precision (64-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[63:0] := tmp_src1[63:0] * POW(2, FLOOR(tmp_src2[63:0]))
	RETURN dst[63:0]
}

dst[63:0] := SCALE(a[63:0], b[63:0])
dst[127:64] := b[127:64]
dst[MAX:128] := 0

Instruction: 'VSCALEFSD'. Intrinsic: '_mm_scalef_sd'. Requires AVX512F.

func ScalefSs 

func ScalefSs(a x86.M128, b x86.M128) (dst x86.M128)

ScalefSs: Scale the packed single-precision (32-bit) floating-point elements in 'a' using values from 'b', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

SCALE(src1, src2){
	IF (src2 == NaN)
		IF (src2 == SNaN)
			RETURN QNAN(src2)
		FI
	ELSE IF (src1 == NaN)
		IF (src1 == SNaN)
			RETURN QNAN(src1)
		FI
		IF (src2 != INF)
			RETURN QNAN(src1)
		FI
	ELSE
		tmp_src2 := src2
		tmp_src1 := src1
		IF (src2 is denormal AND MXCSR.DAZ)
			tmp_src2 := 0
		FI
		IF (src1 is denormal AND MXCSR.DAZ)
			tmp_src1 := 0
		FI
	FI
	dst[31:0] := tmp_src1[31:0] * POW(2, FLOOR(tmp_src2[31:0]))
	RETURN dst[63:0]
}

dst[31:0] := SCALE(a[31:0], b[31:0])
dst[127:32] := b[127:32]
dst[MAX:128] := 0

Instruction: 'VSCALEFSS'. Intrinsic: '_mm_scalef_ss'. Requires AVX512F.

func SqrtRoundSd 

func SqrtRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

SqrtRoundSd: Compute the square root of the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper element from 'b' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := SQRT(a[63:0])
		dst[127:64] := b[127:64]
		dst[MAX:128] := 0

Instruction: 'VSQRTSD'. Intrinsic: '_mm_sqrt_round_sd'. Requires AVX512F.

func SqrtRoundSs 

func SqrtRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

SqrtRoundSs: Compute the square root of the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'b' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := SQRT(a[31:0])
		dst[127:32] := b[127:32]
		dst[MAX:128] := 0

Instruction: 'VSQRTSS'. Intrinsic: '_mm_sqrt_round_ss'. Requires AVX512F.

func SraEpi64 

func SraEpi64(a x86.M128i, count x86.M128i) (dst x86.M128i)

SraEpi64: Shift packed 64-bit integers in 'a' right by 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF count[63:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] >> count[63:0])
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_sra_epi64'. Requires AVX512F.

func SraiEpi64 

func SraiEpi64(a x86.M128i, imm8 byte) (dst x86.M128i)

SraiEpi64: Shift packed 64-bit integers in 'a' right by 'imm8' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	IF imm8[7:0] > 63
		dst[i+63:i] := SignBit
	ELSE
		dst[i+63:i] := SignExtend(a[i+63:i] << imm8[7:0])
	FI
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAQ'. Intrinsic: '_mm_srai_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func SravEpi64 

func SravEpi64(a x86.M128i, count x86.M128i) (dst x86.M128i)

SravEpi64: Shift packed 64-bit integers in 'a' right by the amount specified by the corresponding element in 'count' while shifting in sign bits, and store the results in 'dst'. 

FOR j := 0 to 1
	i := j*64
	dst[i+63:i] := SignExtend(a[i+63:i] >> count[i+63:i])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPSRAVQ'. Intrinsic: '_mm_srav_epi64'. Requires AVX512F.

func SubRoundSd 

func SubRoundSd(a x86.M128d, b x86.M128d, rounding int) (dst x86.M128d)

SubRoundSd: Subtract the lower double-precision (64-bit) floating-point element in 'b' from the lower double-precision (64-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper element from 'a' to the upper element of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[63:0] := a[63:0] - b[63:0]
		dst[127:64] := a[127:64]
		dst[MAX:128] := 0

Instruction: 'VSUBSD'. Intrinsic: '_mm_sub_round_sd'. Requires AVX512F.

func SubRoundSs 

func SubRoundSs(a x86.M128, b x86.M128, rounding int) (dst x86.M128)

SubRoundSs: Subtract the lower single-precision (32-bit) floating-point element in 'b' from the lower single-precision (32-bit) floating-point element in 'a', store the result in the lower element of 'dst', and copy the upper 3 packed elements from 'a' to the upper elements of 'dst'. 

Rounding is done according to the 'rounding' parameter, which can be one

of: 

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		dst[31:0] := a[31:0] - b[31:0]
		dst[127:32] := a[127:32]
		dst[MAX:128] := 0

Instruction: 'VSUBSS'. Intrinsic: '_mm_sub_round_ss'. Requires AVX512F.

func TernarylogicEpi32 

func TernarylogicEpi32(a x86.M128i, b x86.M128i, c x86.M128i, imm8 byte) (dst x86.M128i)

TernarylogicEpi32: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 32-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 3
	i := j*32
	FOR h := 0 to 31
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGD'. Intrinsic: '_mm_ternarylogic_epi32'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func TernarylogicEpi64 

func TernarylogicEpi64(a x86.M128i, b x86.M128i, c x86.M128i, imm8 byte) (dst x86.M128i)

TernarylogicEpi64: Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in 'imm8'. For each bit in each packed 64-bit integer, the corresponding bit from 'a', 'b', and 'c' are used to form a 3 bit index into 'imm8', and the value at that bit in 'imm8' is written to the corresponding bit in 'dst'. 

FOR j := 0 to 1
	i := j*64
	FOR h := 0 to 63
		index[2:0] := (a[i+h] << 2) OR (b[i+h] << 1) OR c[i+h]
		dst[i+h] := imm8[index[2:0]]
	ENDFOR
ENDFOR
dst[MAX:128] := 0

Instruction: 'VPTERNLOGQ'. Intrinsic: '_mm_ternarylogic_epi64'. Requires AVX512F.

FIXME: Requires compiler support (has immediate)

func TestEpi32Mask 

func TestEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

TestEpi32Mask: Compute the bitwise AND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is non-zero. 

FOR j := 0 to 3
	i := j*32
	k[j] := ((a[i+31:i] AND b[i+31:i]) != 0) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTMD'. Intrinsic: '_mm_test_epi32_mask'. Requires AVX512F.

func TestEpi64Mask 

func TestEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

TestEpi64Mask: Compute the bitwise AND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is non-zero. 

FOR j := 0 to 1
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) != 0) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPTESTMQ'. Intrinsic: '_mm_test_epi64_mask'. Requires AVX512F.

func TestnEpi32Mask 

func TestnEpi32Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

TestnEpi32Mask: Compute the bitwise NAND of packed 32-bit integers in 'a' and 'b', producing intermediate 32-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 3
	i := j*32
	k[j] := ((a[i+31:i] NAND b[i+31:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:4] := 0

Instruction: 'VPTESTNMD'. Intrinsic: '_mm_testn_epi32_mask'. Requires AVX512F.

func TestnEpi64Mask 

func TestnEpi64Mask(a x86.M128i, b x86.M128i) (dst x86.Mmask8)

TestnEpi64Mask: Compute the bitwise NAND of packed 64-bit integers in 'a' and 'b', producing intermediate 64-bit values, and set the corresponding bit in result mask 'k' if the intermediate value is zero. 

FOR j := 0 to 1
	i := j*64
	k[j] := ((a[i+63:i] AND b[i+63:i]) == 0) ? 1 : 0
ENDFOR
k[MAX:2] := 0

Instruction: 'VPTESTNMQ'. Intrinsic: '_mm_testn_epi64_mask'. Requires AVX512F.

Types

This section is empty.

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.