jpountz commented on PR #14735: URL: https://github.com/apache/lucene/pull/14735#issuecomment-2924481149
Thank you for looking into this, I remember trying this sort of things out but results were a bit mixed. Your idea of doing a single block at first before moving to more aggressive stepping is interesting. > I modify the code and ran luceneutil, but the result is not good as AdvanceBenchmark showed I had the same problem when working on this optimization, let's not trust AdvanceBenchmark too much, it's hard to re-create similar situation as what actually happens on real data. While your luceneutil benchmarks report some speedups, none of them have a low p-value (the last column), suggesting that you only ran a few iterations. Could you run 20 iterations to help distinguish signal from noise? -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
