luyuncheng opened a new pull request, #12723:
URL: https://github.com/apache/lucene/pull/12723
### Description
I see some hot_thread like following stack,
```
java.lang.Thread.State: RUNNABLE
at org.apache.lucene.store.DataInput.readVInt(DataInput.java:112)
at
org.apache.lucene.util.bkd.BKDReader$BKDPointTree.readDocIDs(BKDReader.java:636)
at
org.apache.lucene.util.bkd.BKDReader$BKDPointTree.visitDocValues(BKDReader.java:608)
at
org.apache.lucene.util.bkd.BKDReader$BKDPointTree.visitLeavesOneByOne(BKDReader.java:595)
at
org.apache.lucene.util.bkd.BKDReader$BKDPointTree.visitDocValues(BKDReader.java:589)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:364)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:357)
at
org.apache.lucene.index.PointValues.intersect(PointValues.java:337)
at
org.apache.lucene.document.SpatialQuery$RelationScorerSupplier.getSparseScorer(SpatialQuery.java:340)
at
org.apache.lucene.document.SpatialQuery$RelationScorerSupplier.getScorer(SpatialQuery.java:306)
at
org.apache.lucene.document.SpatialQuery$2.get(SpatialQuery.java:203)
at org.apache.lucene.search.Weight.bulkScorer(Weight.java:175)
at
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:724)
```
and when we call
`BKDPointTree#visitDocValues#readDocIDs`->`DocIdWriter#readInts`
https://github.com/apache/lucene/blob/c701a5d9be4d8dc677ea994f8ef5fdd8945760d6/lucene/core/src/java/org/apache/lucene/util/bkd/DocIdsWriter.java#L165-L174
it would do iteration to copy docids into `int[] docIDs` when type is
`BITSET_IDS` or `CONTINUOUS_IDS`, and After this, it do iteration again to get
docvalues in `visitDocValuesWithCardinality`
----
### Proposal
i think we can only do only one iteration like this pr shows. AND it can
also reuse `visit(DocIdSetIterator iterator)` optimize when type is
`DocBaseBitSetIterator`
i do benchmark with luceneutil `IndexAndSearchShapes` tests, use params
`-intersects -point -file osmdata.wkt`
it shows:
baseline: BEST QPS: 282.4637999101765
candidate: BEST QPS: 316.8185015110912
--
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
