[libcxx-commits] [libcxx] [libc++] Simplify the implementation of std::sort a bit (PR #104902)

[Alexander Kornienko via libcxx-commits]

alexfh wrote:

In addition to @joanahalili's comment: the regression varies from 5% to 25% depending on the platform. And there's also a much more significant improvement for larger vector sizes (5000+) - around 1.3-2x speedup, but the code is used with smaller vector sizes in prod, which makes the improvement irrelevant to us.

https://github.com/llvm/llvm-project/pull/104902