【发起分享提案】分析 Eisel-Lemire 算法

[yangwenmai]

为什么 Go 1.16 ParseFloat 可以获得性能提升

分析 Eisel-Lemire 算法，了解其根本原因。

Passing "go test -test.count=5 -test.run=xxx -test.bench=Atof64" on to
benchstat:

name                  old time/op  new time/op  delta
Atof64Decimal-4       47.9ns ± 0%  48.3ns ± 1%     ~     (p=0.238 n=4+5)
Atof64Float-4         58.3ns ± 3%  57.7ns ± 0%     ~     (p=0.151 n=5+5)
Atof64FloatExp-4       107ns ± 0%    71ns ± 1%  -33.89%  (p=0.016 n=4+5)
Atof64Big-4            163ns ± 0%   166ns ± 2%     ~     (p=0.159 n=4+5)
Atof64RandomBits-4     299ns ± 1%   166ns ± 1%  -44.41%  (p=0.008 n=5+5)
Atof64RandomFloats-4   188ns ± 1%   144ns ± 0%  -23.03%  (p=0.008 n=5+5)

This does for ParseFloat(etc, 32) what commit a2eb53c571 did for
ParseFloat(etc, 64).

name              old time/op  new time/op  delta
Atof32Decimal-4   48.3ns ± 4%  48.8ns ± 2%     ~     (p=0.548 n=5+5)
Atof32Float-4     56.2ns ± 5%  54.7ns ± 3%     ~     (p=0.246 n=5+5)
Atof32FloatExp-4   104ns ± 0%    76ns ± 2%  -27.19%  (p=0.008 n=5+5)
Atof32Random-4     142ns ± 2%   109ns ± 1%  -23.07%  (p=0.008 n=5+5)

诉求

学习 Go 官方包中采用特定 Eisel-Lemire 算法优化性能。

参考资料

https://go-review.googlesource.com/c/go/+/260858
https://go-review.googlesource.com/c/go/+/264517
https://nigeltao.github.io/blog/2020/eisel-lemire.html

---

备注

任何你想要表达的话语。