msokolov commented on code in PR #14674:
URL: https://github.com/apache/lucene/pull/14674#discussion_r2112296471
##########
lucene/core/src/java/org/apache/lucene/search/AbstractKnnVectorQuery.java:
##########
@@ -226,15 +227,25 @@ private BitSet createBitSet(DocIdSetIterator iterator,
Bits liveDocs, int maxDoc
// If we already have a BitSet and no deletions, reuse the BitSet
return bitSetIterator.getBitSet();
} else {
- // Create a new BitSet from matching and live docs
- FilteredDocIdSetIterator filterIterator =
- new FilteredDocIdSetIterator(iterator) {
- @Override
- protected boolean match(int doc) {
- return liveDocs == null || liveDocs.get(doc);
- }
- };
- return BitSet.of(filterIterator, maxDoc);
+ int threshold = maxDoc >> 7; // same as BitSet#of
+ if (iterator.cost() >= threshold) {
+ // take advantage of Disi#intoBitset and Bits#applyMask
+ FixedBitSet bitSet = new FixedBitSet(maxDoc);
+ bitSet.or(iterator);
+ if (liveDocs != null) {
+ liveDocs.applyMask(bitSet, 0);
+ }
+ return bitSet;
+ } else {
+ FilteredDocIdSetIterator filterIterator =
+ new FilteredDocIdSetIterator(iterator) {
+ @Override
+ protected boolean match(int doc) {
+ return liveDocs == null || liveDocs.get(doc);
Review Comment:
if `liveDocs == null` we could return `Bitset.of(iterator, maxDoc)`, no, and
avoid the internal logic in the filtered iterator, which probably goes away
with branch prediction :shrug:
--
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
