mikemccand commented on PR #14963: URL: https://github.com/apache/lucene/pull/14963#issuecomment-3320914462
These look like amazing results! It's surprising that we see such a gain in indexing throughput? Maybe `knnPerfTest.py` in luceneutil is waiting for all merges to finish in the end? But, I thought we recently added an option to not wait. It might also be that faster merging (not creating HNSW graphs for smallish segments) means more CPU to spend indexing instead of merging, or possibly it changes the backpressure (Lucene will pause indexing threads if merging is falling behind because it sees the indexing threads as evil since they keep producing new segments). Are there any issues we still need to resolve here? -- 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
