[
https://issues.apache.org/jira/browse/LUCENE-10130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17422467#comment-17422467
]
Julie Tibshirani commented on LUCENE-10130:
-------------------------------------------
+1 to switch to a better data structure that could help perf. Some context that
might be helpful: this set tracks the graph nodes that have already been
visited during a kNN search. In
https://issues.apache.org/jira/browse/LUCENE-9937 on dataset of 1M points, I
saw the number of visited nodes range from around 1000 to 20,000, depending on
k. When there are deleted docs this number could climb much higher, since we
might visit a bunch of deleted docs until we find good vectors to return.
> HnswGraph could make use of a SparseFixedBitSet.getAndSet
> ---------------------------------------------------------
>
> Key: LUCENE-10130
> URL: https://issues.apache.org/jira/browse/LUCENE-10130
> Project: Lucene - Core
> Issue Type: Task
> Reporter: Robert Muir
> Priority: Major
> Attachments: LUCENE-10130.patch
>
>
> Currently HnswGraph uses SparseFixedBitSet "visited" to track where it has
> already been. The logic currently looks like this:
> {code}
> if (visited.get(entryPoint) == false) {
> visited.set(entryPoint);
> ... logic ...
> }
> {code}
> If SparseFixedBitSet had a {{getAndSet}} (like FixedBitSet), the code could
> be:
> {code}
> if (visited.getAndSet(entrypoint) == false) {
> ... logic ...
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
