aboutsummaryrefslogtreecommitdiff
path: root/internal/adler32/adler32_generic.go
Commit message (Collapse)AuthorAgeFilesLines
* *: LintGravatar Runxi Yu2026-03-041-0/+5
|
* adler32: ImportGravatar Runxi Yu2026-02-221-0/+44
|
* Delete everything, I'm redesigning this.Gravatar Runxi Yu2026-02-201-45/+0
| | | | | I'll stop using a flat package and make things much more modular. And also experiment with streaming APIs so large blobs don't OOM us.
* adler32: Add package-level doc-commentGravatar Runxi Yu2026-01-301-0/+1
|
* Replace adler32 with github.com/mhr3/adler32-simdGravatar Runxi Yu2025-11-191-75/+23
| | | | Unfortunately no VSX support yet
* SIMD with AVX2 on supported AMD64 machinesGravatar Runxi Yu2025-11-191-0/+96
Some help from gpt-5.1-thinking taken: used wrong register size for the weighted sum at first, so it was truncating the second half of our block; also there was an overflow from the modulus and stuff. Unfortunately the AVX2 adler32 is only about 20% faster than the generic version which doesn't make for much.