Repository navigation

Benchmark array alignment (OBJEXXFCL_ALIGN=64 vs std::vector) with and without native + fast math - #3

Open

jmarrec wants to merge 2 commits into

mainfrom

bench_alignment_vectorization

Open

Benchmark array alignment (OBJEXXFCL_ALIGN=64 vs std::vector) with and without native + fast math#3
jmarrec wants to merge 2 commits into
mainfrom
bench_alignment_vectorization

Commits on Oct 6, 2026
- Initial benchmark for vector alignment, and ffast-math+march/mcpu=native
  jmarrec
  committed
  ffc067b
Commits on Oct 7, 2026
- Alignment bench: OffsetN layouts, non-power-of-two sizes, per-buffer counters, and a script to compare results
  jmarrec
  committed
  686e76d