Minimal proof for Bulk DocIdSetIterator for Lucene PR 13149 #257
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
antonha
commented
Mar 13, 2024
| @@ -422,7 +422,7 @@ def __init__(self, cold=False, | |||
| # Pass fixed randomSeed so separate runs are comparable (pick the same tasks): | |||
| randomSeed=None, | |||
| benchSearch=True, | |||
| taskCountPerCat = 1, | |||
| taskCountPerCat = 20, | |||
There was a problem hiding this comment.
This is important, to trigger multiple implementations of the IntersectVisitor in PointRangeQueries
There was a problem hiding this comment.
I should also note that this might make the benchmark slower, since more implementations of the IntersectVisitor might drag down performance due do virtual calls now being used.
In real Lucene applications these multiple implementations is probably the norm though, so that makes the benchmark better. apache/lucene#13149 should lower the performance decrease from this.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR is the smallest I could make (except for number of LongNrq queries, could probably be fewer) to prove that the changes in apache/lucene/pull/13149 work.
I aimed at reproducing for
wikimediumall. This needs to be run withoptimize = Truefor indexing andcommitPoint = 'single'for the competition - otherwise it is hard to see the performance difference. The reason for this is that the BkdTree IntsWriter otherwise chooses too good of a compression, since the number of documents is too low in each segment.I'm not sure if this should be merged - the PR is mostly here for reference.