MaxGekk commented on a change in pull request #27078: [SPARK-30409][SQL][TESTS] Use `NoOp` datasource in SQL benchmarks URL: https://github.com/apache/spark/pull/27078#discussion_r364572908
########## File path: sql/core/benchmarks/AggregateBenchmark-results.txt ########## @@ -2,142 +2,142 @@ aggregate without grouping ================================================================================================ -OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64 -Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +OpenJDK 64-Bit Server VM 1.8.0_232-8u232-b09-0ubuntu1~18.04.1-b09 on Linux 4.15.0-1051-aws +Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz agg w/o group: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -agg w/o group wholestage off 50499 52524 2863 41.5 24.1 1.0X -agg w/o group wholestage on 1163 1205 56 1803.1 0.6 43.4X +agg w/o group wholestage off 39763 42107 NaN 52.7 19.0 1.0X +agg w/o group wholestage on 967 970 3 2168.7 0.5 41.1X ================================================================================================ stat functions ================================================================================================ -OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64 -Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +OpenJDK 64-Bit Server VM 1.8.0_232-8u232-b09-0ubuntu1~18.04.1-b09 on Linux 4.15.0-1051-aws +Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz stddev: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -stddev wholestage off 13393 13403 14 7.8 127.7 1.0X -stddev wholestage on 1291 1315 25 81.2 12.3 10.4X +stddev wholestage off 6232 6243 15 16.8 59.4 1.0X Review comment: If we look at `Per Row(ns)`, results are comparable: - `stddev wholestage off ` - OpenJDK 8 = 59.4 ns, Oracle JDK 11 = 65.0 ns - `stddev wholestage on` - OpenJDK 8 = 9.1 ns, Oracle JDK 11 = 9.2 ns Comparing to previous runs on OpenJDK 8 doesn't make sense due to some many differences - hardware, JDK version, libs, Linux kernel and so on. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
