9prady9 commented on pull request #11674: URL: https://github.com/apache/arrow/pull/11674#issuecomment-971576899
> That said, I agree with @lidavidm : this may not be worth it. I experimented locally and it's not easy to do this without making performance suffer. My original intuition was that the CPU could easily parallelize popcounts with the load/stores and logical operations, but that seems to depend on compiler optimizations quite a bit. True, unless pop count on a byte is close to a noop (via some special instruction), popcount along with bitmap will incur additional cost. Having said that, since validity/null count on output is only provided on explicit user request, wouldn't it be nice to have the extra feature ? one less API call for user to get pop count on output. I assuming these functions are user facing(as they are exported), hence my comment. -- 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]
