viliam-durina commented on PR #14016: URL: https://github.com/apache/lucene/pull/14016#issuecomment-2516969696
>Won't it change the execution plan used during indexing? As far as I'm aware, it doesn't. >Or -- maybe it would actually be a no-op because the IndexWriter creates a unique segment for each of its threads, so indexing a single segment is always single-threaded. Creating a new segment is always single-threaded in the current version. It can be parallel when merging segments - to enable it, you need to pass a value >=1 to `numMergeWorkers` argument of `Lucene99HnswVectorsFormat`. In my own code I'm implementing concurrent graph building when initially adding the documents. For this, I needed the `MergeSearcher` outside of the context of merging - the locking should ideally be a business of the builder, not of the merger, so I thought the move is logical in general, so I created this PR. Checked the existing tests, I didn't find a test that would test the concurrent merging. However I tested it in my code, and it seems to work... -- 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
