GovindBalaji-S-Glean commented on PR #16051: URL: https://github.com/apache/lucene/pull/16051#issuecomment-6078136477
> Hmm is this because `PrefixCodedTerms` cannot skip (`seekCeil`, `seekFloor`, etc.)? So it's a linear scan of all terms in the query, for each segment? Yes, the decoding work being done for every term in the query for each segment is pretty expensive when there are a lot of terms in the query. I have this old annotated (including unrelated annotations) flame graph of cpu profiles: <img width="370" height="216" alt="image" src="https://github.com/user-attachments/assets/7fb21bbf-f489-4242-a998-8b013b56969a"; /> However, we are not doing any skip in the proposed array implementation, and is much better while still scanning linearly. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
