richardstartin edited a comment on pull request #7920: URL: https://github.com/apache/pinot/pull/7920#issuecomment-997466511
I added a benchmark to show where the difference comes from, but the benchmark also hints at the minimum cost of query vectorization: an array copy. This cost could be eliminated for reductions. ``` Benchmark (_blockSize) (_numBlocks) Mode Cnt Score Error Units BenchmarkFixedByteSVForwardIndexReader.readDoubles 10000 1000 avgt 5 37.284 ± 0.919 ms/op <- bad, long to double conversion BenchmarkFixedByteSVForwardIndexReader.readDoublesBatch 10000 1000 avgt 5 15.674 ± 0.245 ms/op <- no bounds checks BenchmarkFixedByteSVForwardIndexReader.readLongs 10000 1000 avgt 5 35.244 ± 1.947 ms/op <- bad, but no type conversion BenchmarkFixedByteSVForwardIndexReader.readLongsBatch 10000 1000 avgt 5 10.777 ± 0.163 ms/op <- best case, vectorized copy, no type conversion ``` -- 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: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
