Skip to content

perf: adds AVX2 optimization for WebSocket masking#554

Open
DarkGL wants to merge 3 commits intocoder:masterfrom
DarkGL:perf/optimize-mask
Open

perf: adds AVX2 optimization for WebSocket masking#554
DarkGL wants to merge 3 commits intocoder:masterfrom
DarkGL:perf/optimize-mask

Conversation

@DarkGL
Copy link

@DarkGL DarkGL commented Jan 12, 2026

Before

Benchmark_mask/amd64/wdvxdr1123-asm/8-12        444599608                2.711 ns/op    2951.47 MB/s           0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/16-12               414228843                2.908 ns/op    5501.26 MB/s           0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/32-12               405062330                2.916 ns/op    10974.92 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/128-12              323147343                3.702 ns/op    34575.47 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/256-12              181181568                6.595 ns/op    38815.81 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/512-12              134261024                8.932 ns/op    57323.30 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/1024-12             87677349                13.57 ns/op     75469.43 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/2048-12             49833375                24.09 ns/op     85028.41 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/4096-12             24303784                49.48 ns/op     82783.45 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/8192-12             12023418               100.2 ns/op      81742.79 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/16384-12             6031188               197.7 ns/op      82860.28 MB/s          0 B/op          0 allocs/op

After

Benchmark_mask/amd64/wdvxdr1123-asm/8-12        419437528                2.896 ns/op    2762.68 MB/s           0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/16-12               365261996                3.265 ns/op    4900.24 MB/s           0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/32-12               365955547                3.274 ns/op    9775.18 MB/s           0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/128-12              295290266                4.057 ns/op    31550.39 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/256-12              293986858                4.070 ns/op    62905.39 MB/s          0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/512-12              244795381                4.893 ns/op    104638.32 MB/s         0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/1024-12             162367981                7.413 ns/op    138135.75 MB/s         0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/2048-12             89116902                13.52 ns/op     151521.17 MB/s         0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/4096-12             46351291                25.72 ns/op     159256.73 MB/s         0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/8192-12             23834427                50.61 ns/op     161874.64 MB/s         0 B/op          0 allocs/op
Benchmark_mask/amd64/wdvxdr1123-asm/16384-12            12019286                99.17 ns/op     165210.00 MB/s         0 B/op          0 allocs/op

@DarkGL DarkGL marked this pull request as draft January 12, 2026 14:45
@DarkGL DarkGL marked this pull request as ready for review January 12, 2026 14:55
@DarkGL
Copy link
Author

DarkGL commented Jan 13, 2026

@mafredri is it possible to rerun test ? I can't reproduce fail on localhost

@mafredri
Copy link
Member

mafredri commented Jan 22, 2026

@DarkGL I haven't had time to check your PR yet, but thank you! I just started a rerun of CI, FWIW it looks like an ephemeral issue I've seen before.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@DarkGL @mafredri