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

[Louis Dionne via libcxx-commits]

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

----------------
ldionne wrote:

The idea behind these benchmarks was that we were batching the operations exactly so we could then afford to Pause and Resume. If we don't do that, then we're basically benchmarking both `push_back` and `erase`/`clear`, not only `push_back`. I think we need to address this differently.

Maybe we could have 10-20 vectors that we all `reserve` and then `push_back` into, *and then* we pause, erase, and resume?

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