~runxiyu/furgit.git/internal/adler32/bench_test.go, branch v0.1.1 Git library in pure Go https://cgit.space/~runxiyu/furgit.git/atom/internal/adler32/bench_test.go?h=v0.1.1 2025-11-20T00:00:00Z adler32: Add benchmark 2025-11-20T00:00:00Z Runxi Yu 2025-11-20T00:00:00Z urn:sha1:2d70cfd89c2a3ab5af00ba8bb557a36ef8395a02 Replace adler32 with github.com/mhr3/adler32-simd 2025-11-19T00:00:03Z Runxi Yu 2025-11-19T00:00:00Z urn:sha1:7d60b65249c8af355d739ccf44482a899d89c9c1 Unfortunately no VSX support yet SIMD with AVX2 on supported AMD64 machines 2025-11-19T00:00:00Z Runxi Yu 2025-11-19T00:00:00Z urn:sha1:676f08cc4b0deb397033598ac96c0d0272fcae05 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.