Files
crypto/blake2s
Xiaolin Zhao 2ebaafcdf5 blake2s: add loong64 SIMD implementation
The performance gains on Loongson 3A6000 and 3A5000 are as follows:

goos: linux
goarch: loong64
pkg: golang.org/x/crypto/blake2s
cpu: Loongson-3A6000-HV @ 2500.00MHz
        |  bench.old  |              bench.new              |
        |   sec/op    |   sec/op     vs base                |
Write64   277.8n ± 0%   113.5n ± 0%  -59.14% (p=0.000 n=10)
Write1K   4.246µ ± 0%   1.736µ ± 0%  -59.11% (p=0.000 n=10)
Sum64     289.9n ± 0%   137.7n ± 0%  -52.51% (p=0.000 n=10)
Sum1K     4.265µ ± 0%   1.758µ ± 0%  -58.78% (p=0.000 n=10)
geomean   1.099µ        467.3n       -57.48%

        |  bench.old   |               bench.new               |
        |     B/s      |     B/s       vs base                 |
Write64   219.7Mi ± 0%   537.9Mi ± 0%  +144.86% (p=0.000 n=10)
Write1K   230.0Mi ± 0%   562.6Mi ± 0%  +144.62% (p=0.000 n=10)
Sum64     210.5Mi ± 0%   443.3Mi ± 0%  +110.59% (p=0.000 n=10)
Sum1K     229.0Mi ± 0%   555.5Mi ± 0%  +142.64% (p=0.000 n=10)
geomean   222.1Mi        522.5Mi       +135.21%

goos: linux
goarch: loong64
pkg: golang.org/x/crypto/blake2s
cpu: Loongson-3A5000 @ 2500.00MHz
        |  bench.old  |              bench.new              |
        |   sec/op    |   sec/op     vs base                |
Write64   373.8n ± 0%   175.0n ± 0%  -53.18% (p=0.000 n=10)
Write1K   5.763µ ± 0%   2.595µ ± 0%  -54.97% (p=0.000 n=10)
Sum64     397.8n ± 0%   205.7n ± 0%  -48.29% (p=0.000 n=10)
Sum1K     5.787µ ± 0%   2.627µ ± 0%  -54.61% (p=0.000 n=10)
geomean   1.492µ        703.8n       -52.83%

        |  bench.old   |               bench.new               |
        |     B/s      |     B/s       vs base                 |
Write64   163.3Mi ± 0%   348.9Mi ± 0%  +113.62% (p=0.000 n=10)
Write1K   169.5Mi ± 0%   376.3Mi ± 0%  +122.09% (p=0.000 n=10)
Sum64     153.4Mi ± 0%   296.7Mi ± 0%   +93.37% (p=0.000 n=10)
Sum1K     168.7Mi ± 0%   371.8Mi ± 0%  +120.33% (p=0.000 n=10)
geomean   163.6Mi        346.9Mi       +112.03%

Change-Id: Id91ffbefc538bce294875d72e6cde72fea43afbf
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/661215
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Carlos Amedee <carlos@golang.org>
Auto-Submit: Carlos Amedee <carlos@golang.org>
Reviewed-by: abner chenc <chenguoqi@loongson.cn>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
2025-04-11 13:55:16 -07:00
..
2023-10-11 22:02:22 +00:00
2018-10-29 02:12:03 +00:00