float16

Documentation

Index

• func Float16ToUint64(v uint16) uint64
• func Uint64ToFloat16RoundUp(v uint64) (uint16, bool)

Functions

func Float16ToUint64

func Float16ToUint64(v uint16) uint64

Float16ToUint64 converts a 16-bit floating-point value to its corresponding integer value. This function is the inverse of Uint64ToFloat16RoundUp.

func Uint64ToFloat16RoundUp

func Uint64ToFloat16RoundUp(v uint64) (uint16, bool)

Uint64ToFloat16RoundUp converts an integer value to its corresponding 16-bit floating-point value, rounding up if needed.

Note that the format used differs from IEEE 754's half-precision floating-point format. There is no sign bit, meaning that the significand is 11 bits. The exponent bias is chosen so that the smallest subnormal value corresponds to integer value 1. There is no support for expressing infinity/NaN.