iprithv commented on PR #16047: URL: https://github.com/apache/lucene/pull/16047#issuecomment-4446873781
@benwtrent Thanks for the pointer to #16034, agreed it's closer to what the issue describes, since sibling scoring happens inline during HNSW traversal and can affect early termination, while this POC only rescore parents after HNSW finishes (so it can fix parent score correctness but doesn't change traversal). A couple of clarifying questions before I close this out: 1. Is there still value in a much simpler post-HNSW path (no core API changes, no docId→ordinal cache) as a lower-overhead option for users who want correct parent ranking but not the recall via early termination story? Or do you see that as too narrow to be worth a separate code path? 2. Wanted to check if anything from the JMH harness here (DiversifyingChildrenFloatKnnJoinBenchmark) worth porting over? Thanks! -- 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]
