[libc-commits] [libc] [libc] Run Benchmarks on 32 Threads by Default (PR #98964)

[Joseph Huber via libc-commits]

https://github.com/jhuber6 commented:

Hmm, I feel like it's easy enough to just set this one each one. It's also difficult because the warp size varies on the hardware (and compilation settings) for AMDGPU. Having a helper for a single threaded run is probably fine.

https://github.com/llvm/llvm-project/pull/98964