etseidl commented on PR #8602: URL: https://github.com/apache/arrow-rs/pull/8602#issuecomment-3444901295
I think I want to start on a parquet end-to-end benchmark of some sort. The arrow_reader bench is *so* sensitive. On my workstation I was seeing the BinaryViewArray bench taking about 10% longer consistently across the various permutations. Then I merged the latest from main, and one bench regressed further, but the others were consistently faster than baseline. I then reverted one change in this PR...the first regression nearly went away, but the others then regressed significantly. 🤯 I think I'm convinced these changes won't have much impact on performance. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
