navneet1v commented on PR #15722: URL: https://github.com/apache/lucene/pull/15722#issuecomment-3928725196
> We need to benchmark with quantized vectors, where the vector ops during traversal are magnitudes cheaper, not just floating point. I am working on the benchmarks. > Prefetch larger blocks hoping to get many vectors we care about besides the immediate ordinals. This is basic implementation I am thinking to combine the few ordinals in a block say 128KB(configurable) to do prefetch. Will it be better if I remove the prefetch from scorer and graph searcher and move it to only KNNVectorValues. So that at-least for rescoring or for their own scorer prefetch can be used. -- 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]
