command
module
Version:
v0.0.0-...-faa7dad
Opens a new window with list of versions in this module.
Published: Jun 11, 2018
License: MIT
Opens a new window with license information.
Imports: 2
Opens a new window with list of imports.
Imported by: 0
Opens a new window with list of known importers.
README
¶
ELTEE
(LT) ELTEE stands for "LOADTEST"
Benchmarks
Running all the beanchmaks at once
$ go test -bench=. -v ./...
Some results
BenchmarkIntChannelAppend-8 50000000 26.6 ns/op
BenchmarkIntChannelDrain-8 30000000 53.7 ns/op
BenchmarkIntSliceAppend-8 300000000 3.91 ns/op
BenchmarkIntMapAppend-8 10000000 151 ns/op
BenchmarkIntSyncMapAppend-8 20000000 106 ns/op
BenchmarkIntSyncMapAppendParallel-8 10000000 198 ns/op
BenchmarkGoNumStdDev5000-8 100000 12121 ns/op
BenchmarkGoNumMean5000-8 200000 6127 ns/op
BenchmarkMathMin5000-8 100000 22812 ns/op
BenchmarkMathMax5000-8 100000 23532 ns/op
BenchmarkIntChannelParallelAppend-8 30000000 49.2 ns/op
Insights from the benchmarks
- Slices are fast
- Map are sinigicantly slower than slices (around 7 times slower)
- With increasing number os keys, maps tend not to change acess time too much for same amount of keys
- Key length and type matters to maps because of the hashing function used internally
- Map from sync.Map is a somewhat good option when parallel write access is required
- Channels are surprisingly fast for both append and drains
- Math functions can we quite slow for specifically Min & Max are substancially slower than first imagined
Documentation
¶
There is no documentation for this package.
Source Files
¶
Directories
¶
Click to show internal directories.
Click to hide internal directories.