costin commented on PR #16141: URL: https://github.com/apache/lucene/pull/16141#issuecomment-4563904223
I reran `DefaultBulkScorerBenchmark` locally on the latest PR branch on my laptop (so there's some noise) after removing the runtime `scanIsEmpty()` calls. While the raw numbers are different than on the original post, the relative speedup remains. Command: `java -jar lucene/benchmark-jmh/build/benchmarks/lucene-benchmark-jmh-11.0.0-SNAPSHOT.jar org.apache.lucene.benchmark.jmh.DefaultBulkScorerBenchmark -rf json -rff /tmp/default-bulkscorer-benchmark.json` JDK: Temurin 25.0.3. Higher throughput is better. | density | deletions | baseline ops/ms | intoBitSet ops/ms | speedup | |---:|:---:|---:|---:|---:| | 0.01 | false | 47.112 | 83.921 | 1.8x | | 0.01 | true | 29.248 | 53.997 | 1.8x | | 0.1 | false | 3.081 | 66.020 | 21.4x | | 0.1 | true | 2.472 | 45.431 | 18.4x | | 0.5 | false | 0.537 | 84.557 | 157.6x | | 0.5 | true | 0.444 | 60.437 | 136.2x | | 1.0 | false | 0.232 | 404.220 | 1745.1x | | 1.0 | true | 0.227 | 3.091 | 13.6x | | -- 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]
