ChrisHegarty commented on PR #12703: URL: https://github.com/apache/lucene/pull/12703#issuecomment-1772963134
@rmuir If I understand your comment correctly. I unaligned the vector data in the mmap file, in the benchmark. The results are similar enough to the aligned, maybe a little less when the alignment is terrible! see [b02b79b](https://github.com/apache/lucene/pull/12703/commits/b02b79b0b3aafdb889a5314dc07a235c0eac45db) Rocket Lake, AVX 512 ``` INFO: Java vector incubator API enabled; uses preferredBitSize=512 ... VectorUtilBenchmark.floatDotProductVector 1024 thrpt 5 9.994 ± 0.002 ops/us VectorUtilBenchmark.floatDotProductVectorMS1 1024 thrpt 5 11.171 ± 0.007 ops/us VectorUtilBenchmark.floatDotProductVectorMS2 1024 thrpt 5 13.159 ± 0.017 ops/us ``` Mac M2, 128 bit ``` INFO: Java vector incubator API enabled; uses preferredBitSize=128 ... VectorUtilBenchmark.floatDotProductVector 1024 thrpt 5 5.242 ± 0.025 ops/us VectorUtilBenchmark.floatDotProductVectorMS1 1024 thrpt 5 7.928 ± 0.155 ops/us VectorUtilBenchmark.floatDotProductVectorMS2 1024 thrpt 5 12.546 ± 0.077 ops/us ``` -- 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
