benwtrent commented on issue #13768: URL: https://github.com/apache/lucene/issues/13768#issuecomment-2344070572
I agree, this is worth digging into. In my benchmarking, `compressed=true` ended up being even marginally faster than `compressed=false`, but I didn't know why :/ which is why I was hesitant to remove the option before. It would be good to benchmark this with and without panamavector enabled over intel & arm. Note, there has been work to move all quantized comparisons off-heap: https://github.com/apache/lucene/pull/13497 the results of which may or may not effect this decision (bytes no longer being copied on heap, thus the fewer bytes being copied no longer harm/help performance). But, I kept having weird slow downs that I cannot figure out and nobody else can replicate (@ChrisHegarty has tried and couldn't see why I keep seeing significant slow downs given different JDK versions). -- 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
