jimczi commented on code in PR #12529:
URL: https://github.com/apache/lucene/pull/12529#discussion_r1319990727
##########
lucene/backward-codecs/src/java/org/apache/lucene/backward_codecs/lucene90/Lucene90HnswVectorsReader.java:
##########
@@ -423,8 +422,12 @@ public RandomAccessVectorValues<float[]> copy() {
@Override
public float[] vectorValue(int targetOrd) throws IOException {
+ if (lastOrd == targetOrd) {
+ return value;
+ }
Review Comment:
Not on normal search where only the query vector is used multiple times but
it happens frequently when we build the index during the diversification phase.
There, we compare single ordinals with all their neighbors sequentially so
instead of having the optimisation in the graph I moved it to the reader since
it's straightforward to add and shouldn't add any overhead to check.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
