Speed up batch scans and Python calls - #53

Merged

mrecachinas merged 9 commits into

mainfrom

perf/theoretical-floor

Jul 10, 2026

Commits on Jul 10, 2026

perf(array): stop best scan at exact match — 99.5% on 512x16

mrecachinas
and
Copilot
committed
b107a5f
perf(array): bound Rayon fan-out — 84.8% on 100000x128 best

mrecachinas
and
Copilot
committed
a6cae2c
perf(python): bypass PyBuffer for bytes — 64.2% on within-distance

mrecachinas
and
Copilot
committed
685a82d
perf(python): keep small buffers attached — 22.7% on bytearray best

mrecachinas
and
Copilot
committed
5badb2c
perf(rust): inline hex entrypoint — 21.8% on 16-char

mrecachinas
and
Copilot
committed
b5e165f
perf(neon): check first hex block eagerly — 49.9% on tight 1024-char

mrecachinas
and
Copilot
committed
0940c2a
perf(neon): scale eager blocks with cutoff — 47.3% on random max100

mrecachinas
and
Copilot
committed
d75a6d6
perf(aarch64): prefer native byte kernel — 24.0% on 64-byte Rust

mrecachinas
and
Copilot
committed
d2a50cb
docs: refresh M4 Max performance benchmarks

mrecachinas
and
Copilot
committed
38420de