Files
Odin/core/math/rand
Yawning Angel a171746a15 core/math/rand: Disable LLVM vectorization of xoshiro on Intel
LLVM tries really hard to vectorize this when it is a terrible idea,
when `-microarch:native` is specified on modern processors.  Not doing
so is a 2.5->3x performance gain.
2026-03-19 19:25:23 +09:00
..