keith-turner commented on PR #31: URL: https://github.com/apache/accumulo-access/pull/31#issuecomment-1802975566
The following is the result of running the benchmark on commit 7ea7b4e from this PR. ``` Benchmark Mode Cnt Score Error Units AccessExpressionBenchmark.measureBytesValidation thrpt 12 29.083 ± 1.099 ops/us AccessExpressionBenchmark.measureEvaluation thrpt 12 12.429 ± 0.280 ops/us AccessExpressionBenchmark.measureStringValidation thrpt 12 24.336 ± 1.423 ops/us ``` The following is the result of running the benchmark on 46b5b877c3eaf433b122812701f47ae5c65ff0c4 from main. ``` Benchmark Mode Cnt Score Error Units AccessExpressionBenchmark.measureBytesParsing thrpt 12 12.597 ± 0.257 ops/us AccessExpressionBenchmark.measureEvaluation thrpt 12 20.407 ± 0.287 ops/us AccessExpressionBenchmark.measureEvaluationAndParsing thrpt 12 8.235 ± 0.068 ops/us AccessExpressionBenchmark.measureStringParsing thrpt 12 11.267 ± 0.285 ops/us ``` In main the `measureBytesParsing` is roughly equivalent to `measureBytesValidation` in this PR. That goes from 12.597 to 29.083 which is a 2.31 times faster. This will be really nice for creating column visibility objects which only need to validate. In main `measureEvaluation` has no equivalent in this PR because that was testing using an existing parse tree, there is no longer a parse tree. This was not a useful benchmark because pre-parsed expression would not be available when streaming through data. In main `measureEvaluationAndParsing` is equivalent to `measureEvaluation` in this PR. That goes from 8.235 to 12.429 which is 1.51 times faster. This will be a nice speedup for evaluating visibility expressions on the server side in Accumulo. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
