mayya-sharipova commented on PR #11743: URL: https://github.com/apache/lucene/pull/11743#issuecomment-1238235494
@msokolov > Although ... the RAM needed for the graph was always required, even when building the graph during flush, it just wasn't accounted for I think. I suppose a possible way to improve the buffering situation would be to buffer the vectors in RAM and then on merge, write them out, freeing the on-heap copy, and while building the graph, access the vectors from disk Indeed, that how we do that in Lucene 9.3, [using off-heap vector values to build graph](https://github.com/apache/lucene/blob/branch_9_3/lucene/core/src/java/org/apache/lucene/codecs/lucene92/Lucene92HnswVectorsWriter.java#L148-L154) The problem with this is that building graph on flush would take a lot of time, which makes searches that needed updated changes unpredictably long. -- 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
