README
¶
Float8
minifloat in Golang
In computing, minifloats are floating-point values represented with very few bits. The library implements float8 (8-bit uint8). It ideal for applications where memory and storage efficiency are crucial but lossy precision is acceptable (e.g. computer graphics, manche learning, etc).
Features
- IEEE 754 and FP8 E4M3 compatible format.
- Fast conversion from/to float32.
- Fast algebraic operations (+, -, *, /).
Getting Started
The latest version of the module is available at main branch. All development, including new features and bug fixes, take place on the main branch using forking and pull requests as described in contribution guidelines. The stable version is available via Golang modules.
Use go get to retrieve the library and add it as dependency to your application.
go get -u github.com/kshard/float8
Quick example
package main
import (
"fmt"
"math"
"github.com/kshard/float8"
)
func main() {
phi08 := float8.ToFloat8(math.Phi)
phi32 := float8.ToFloat32(phi08)
fmt.Printf("𝝅(08) %f : %08b \n", phi32, phi08)
fmt.Printf("𝝅(32) %f : %032b \n", float32(math.Phi), math.Float32bits(math.Phi))
}
The conversion is lossy and supported range of values is limited.
Benchmark
The library implement public api using code books so that its pure Go code is efficient:
go test -run=^$ -bench=. -benchtime=1s -cpu 1
BenchmarkToFloat8 546344320 1.9410 ns/op
BenchmarkToFloat32 1000000000 0.5158 ns/op
BenchmarkAdd 1000000000 0.5804 ns/op
BenchmarkMul 1000000000 0.6461 ns/op
BenchmarkToSlice8 3481468 348.70 ns/op
The internal package math8 implements float-point algebra with focus on correctness, which is used to build code books.
How To Contribute
The library is MIT licensed and accepts contributions via GitHub pull requests:
- Fork it
- Create your feature branch (
git checkout -b my-new-feature) - Commit your changes (
git commit -am 'Added some feature') - Push to the branch (
git push origin my-new-feature) - Create new Pull Request
The build and testing process requires Go version 1.21 or later.
commit message
The commit message helps us to write a good release note, speed-up review process. The message should address two question what changed and why. The project follows the template defined by chapter Contributing to a Project of Git book.
bugs
If you experience any issues with the library, please let us know via GitHub issues. We appreciate detailed and accurate reports that help us to identity and replicate the issue.
License
References
Documentation
¶
Overview ¶
DO NOT EDIT! Use cmd to regenerate it.
DO NOT EDIT! Use cmd to regenerate it.
DO NOT EDIT! Use cmd to regenerate it.
Package float8 implement minifloat (https://en.wikipedia.org/wiki/Minifloat) compatible with IEEE 754 and FP8 E4M3 The number is defined as ±mantissa × 2^exponent
DO NOT EDIT! Use cmd to regenerate it.
DO NOT EDIT! Use cmd to regenerate it.
Index ¶
Constants ¶
const (
Infinity = 0x7f | mantissaMask
)
Variables ¶
This section is empty.
Source Files
Directories