RamakrishnaChilaka opened a new pull request, #15122: URL: https://github.com/apache/lucene/pull/15122
### Description After reviewing the PR: https://github.com/apache/lucene/pull/15065, I considered further improving it by unrolling the for loop. This PR optimizes the performance of VInt and GroupVInt benchmarks by unrolling loops. #### Baseline: ``` Benchmark (size) Mode Cnt Score Error Units GroupVIntBenchmark.benchByteArrayDataInput_readGroupVInt 64 thrpt 5 9.922 ± 0.079 ops/us GroupVIntBenchmark.benchByteArrayDataInput_readVInt 64 thrpt 5 13.411 ± 0.047 ops/us GroupVIntBenchmark.benchByteBuffersIndexInput_readGroupVInt 64 thrpt 5 6.122 ± 0.007 ops/us GroupVIntBenchmark.benchByteBuffersIndexInput_readGroupVIntBaseline 64 thrpt 5 3.300 ± 0.082 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readGroupVInt 64 thrpt 5 14.618 ± 0.011 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readGroupVIntBaseline 64 thrpt 5 11.376 ± 0.369 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readVInt 64 thrpt 5 13.678 ± 0.044 ops/us GroupVIntBenchmark.benchNIOFSDirectoryInputs_readGroupVInt 64 thrpt 5 11.356 ± 0.016 ops/us GroupVIntBenchmark.benchNIOFSDirectoryInputs_readGroupVIntBaseline 64 thrpt 5 8.437 ± 0.039 ops/us GroupVIntBenchmark.bench_writeGroupVInt 64 thrpt 5 5.745 ± 0.208 ops/us VIntBenchmark.benchByteArrayDataInput_readVInt 64 thrpt 5 12.726 ± 0.301 ops/us VIntBenchmark.benchByteArrayDataInput_readVLong 64 thrpt 5 7.748 ± 0.181 ops/us VIntBenchmark.benchMMapDirectoryInputs_readVInt 64 thrpt 5 13.689 ± 0.008 ops/us VIntBenchmark.benchMMapDirectoryInputs_readVLong 64 thrpt 5 6.655 ± 0.031 ops/us ``` #### Contender ``` Benchmark (size) Mode Cnt Score Error Units GroupVIntBenchmark.benchByteArrayDataInput_readGroupVInt 64 thrpt 5 9.963 ± 0.371 ops/us GroupVIntBenchmark.benchByteArrayDataInput_readVInt 64 thrpt 5 19.952 ± 0.538 ops/us GroupVIntBenchmark.benchByteBuffersIndexInput_readGroupVInt 64 thrpt 5 6.128 ± 0.014 ops/us GroupVIntBenchmark.benchByteBuffersIndexInput_readGroupVIntBaseline 64 thrpt 5 3.288 ± 0.039 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readGroupVInt 64 thrpt 5 14.596 ± 0.046 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readGroupVIntBaseline 64 thrpt 5 21.632 ± 0.282 ops/us GroupVIntBenchmark.benchMMapDirectoryInputs_readVInt 64 thrpt 5 16.079 ± 0.097 ops/us GroupVIntBenchmark.benchNIOFSDirectoryInputs_readGroupVInt 64 thrpt 5 11.355 ± 0.026 ops/us GroupVIntBenchmark.benchNIOFSDirectoryInputs_readGroupVIntBaseline 64 thrpt 5 8.598 ± 0.016 ops/us GroupVIntBenchmark.bench_writeGroupVInt 64 thrpt 5 5.758 ± 0.180 ops/us VIntBenchmark.benchByteArrayDataInput_readVInt 64 thrpt 5 19.378 ± 0.038 ops/us VIntBenchmark.benchByteArrayDataInput_readVLong 64 thrpt 5 11.414 ± 0.061 ops/us VIntBenchmark.benchMMapDirectoryInputs_readVInt 64 thrpt 5 17.928 ± 0.132 ops/us VIntBenchmark.benchMMapDirectoryInputs_readVLong 64 thrpt 5 10.053 ± 0.064 ops/us ``` #### Summarising: | Benchmark | Baseline Score (ops/us) | Contender Score (ops/us) | Change (%) | | ------------------------------------------------------------------- | ----------------------- | ------------------------ | ---------- | | GroupVIntBenchmark.benchByteArrayDataInput\_readGroupVInt | 9.922 | 9.963 | +0.41 | | GroupVIntBenchmark.benchByteArrayDataInput\_readVInt | 13.411 | 19.952 | +48.78 | | GroupVIntBenchmark.benchByteBuffersIndexInput\_readGroupVInt | 6.122 | 6.128 | +0.09 | | GroupVIntBenchmark.benchMMapDirectoryInputs\_readGroupVInt | 14.618 | 14.596 | -0.15 | | GroupVIntBenchmark.benchMMapDirectoryInputs\_readGroupVIntBaseline | 11.376 | 21.632 | +89.88 | | GroupVIntBenchmark.benchMMapDirectoryInputs\_readVInt | 13.678 | 16.079 | +17.56 | | GroupVIntBenchmark.benchNIOFSDirectoryInputs\_readGroupVInt | 11.356 | 11.355 | -0.01 | | GroupVIntBenchmark.benchNIOFSDirectoryInputs\_readGroupVIntBaseline | 8.437 | 8.598 | +1.91 | | GroupVIntBenchmark.bench\_writeGroupVInt | 5.745 | 5.758 | +0.23 | | VIntBenchmark.benchByteArrayDataInput\_readVInt | 12.726 | 19.378 | +52.27 | | VIntBenchmark.benchByteArrayDataInput\_readVLong | 7.748 | 11.414 | +47.32 | | VIntBenchmark.benchMMapDirectoryInputs\_readVInt | 13.689 | 17.928 | +30.97 | | VIntBenchmark.benchMMapDirectoryInputs\_readVLong | 6.655 | 10.053 | +51.06 | -- 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
