dweiss commented on issue #14315: URL: https://github.com/apache/lucene/issues/14315#issuecomment-2692034559
Thanks for sharing the kernel performance tracking magic! I'll certainly try it. What is puzzling to me is why - if it's a problem on the java side somewhere - there are such large variations between tmpfs and other filesystems (and between filesystems too). Maybe I'll get to understand it, with some debugging magic of yours. rep_stos_alternative is perhaps not 100% of the flamegraph I posted above but it's certainly a large chunk. I also looked briefly at the code of the higher-up iomap_dio_rw but I don't seem anything particularly suspicious there. I have not looked at the arguments though. I'll keep chipping at this, lovely low-level stuff for a change. -- 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
