[libcxx-commits] [libcxx] [libc++] Remove {Pause, Resume}Timing from fast push_back benchmarks (PR #209130)

[Nikolas Klauser via libcxx-commits]

================

----------------
philnik777 wrote:

We're actually already batching for `push_back() (with reserve)`, and even increasing the batching by 10x still shows significant overhead. Notably, even the cheap element case shows significant overhead in some cases, so `is_trivially_destructible` probably isn't the right condition either. I've added them back for the `many elements` case though, since actually allocating in there makes the overhead mostly irrelevant.

https://github.com/llvm/llvm-project/pull/209130