gf2121 commented on pull request #541: URL: https://github.com/apache/lucene/pull/541#issuecomment-1010252222
Thanks @jpountz ! > I'd be curious to see how it would compare with a similar approach where we would use an int[] to hold the doc IDs instead of longs. I wrote a simple micro benchmark and result looks promising (especially for 32 bit) , I'd like to give it a try. ``` Benchmark (byteOrder) Mode Cnt Score Error Units PackedIntsDecodeBenchmark.decode16ToInt LE thrpt 10 15603376.770 ± 229121.911 ops/s PackedIntsDecodeBenchmark.decode16ToLong LE thrpt 10 11128632.044 ± 1037046.154 ops/s PackedIntsDecodeBenchmark.decode24ToInt LE thrpt 10 9494705.748 ± 1773057.040 ops/s PackedIntsDecodeBenchmark.decode24ToLong LE thrpt 10 7741109.174 ± 838010.286 ops/s PackedIntsDecodeBenchmark.decode32ToInt LE thrpt 10 20356729.998 ± 306557.117 ops/s PackedIntsDecodeBenchmark.decode32ToLong LE thrpt 10 9907810.235 ± 1800010.522 ops/s ``` -- 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
