jpountz commented on issue #12826: URL: https://github.com/apache/lucene/issues/12826#issuecomment-1821390184
I updated the benchmark to interleave numbers of bytes for each integer, plus added a couple minor optimizations to the decoding logic, and I'm now seeing better performance for the new impl, at least on my machine. ``` Benchmark (numBytesPerInt) (size) Mode Cnt Score Error Units GroupVIntBenchmark.byteArrayReadGroupVInt 0 64 thrpt 5 6.434 ± 0.271 ops/us GroupVIntBenchmark.byteBufferReadGroupVInt 0 64 thrpt 5 4.326 ± 0.038 ops/us GroupVIntBenchmark.memorySegmentReadGroupVInt 0 64 thrpt 5 7.913 ± 0.034 ops/us ``` Since this depends on what the actual data looks like, it would be goot to check on a `luceneutil` macrobenchmark, like you did the other day, to check if it's actually faster or not. -- 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
