benwtrent opened a new issue, #15219: URL: https://github.com/apache/lucene/issues/15219
### Description Reading through the code of the BpVectorReorderer, I noticed that it simply reads the vectors on heap and then does individual comparisons between the centroids and the vectors used in the threads. I think it would be much faster to bulk score against the left and right centroids instead. I am not 100% sure if this is directly available over a merged view (that might be step one...). I am particularly looking at the `computeBias` method. Seems like the caller in the `computeBiasTask` could create a leftScorer and a rightScorer and bulk score over each centroid. -- 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.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
