[libc-commits] [libc] [libc] Add Generic and NVPTX Sin Benchmark (PR #99795)

[via libc-commits]

jameshu15869 wrote:

> Just call it once when we initialize the benchmarks, the sequence will then effectively be "random" whenever any other thread gets a value from the sequence.

I'm thinking about having thread 0 call `srand()` right before we start iterating through our benchmark list and running them (So it's only called once throughout the execution) - is that close to what you were thinking?

https://github.com/llvm/llvm-project/pull/99795