Dandandan commented on PR #7463: URL: https://github.com/apache/arrow-rs/pull/7463#issuecomment-2873808297
> I ran [0a01037](https://github.com/apache/arrow-rs/commit/0a01037047dfa4a9425debbcad70e5fd69e24ab1) (= main) against [2f2905e](https://github.com/apache/arrow-rs/commit/2f2905eafec8e924efd846d0340b93cee8de532c) (= branch/PR) in the order main, branch, main, branch: > > and here w/ a 10% threshold: > > It seems like these benchmark are rather noisy and I wonder if instead of wallclock time we should measure instruction counts or something like that. > > That said, I don't think any of them really regress w/ this PR and a few have substantial improvements, so I would :shipit: Thanks for doing the analysis! I agree - the benchmarks should ideally be less noisy. Instruction counts is possible, although it probably misses allocations. My results were a bit different, it seems macOS is giving some different performance results. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
