Github user dbtsai commented on the issue:
https://github.com/apache/spark/pull/22966
cc @jleach4 and @aokolnychyi
We have a great success using
[jmh](http://openjdk.java.net/projects/code-tools/jmh/) for this type of
benchmarking; the benchmarks can be written in the unit test. This framework
handles JVM warn-up, computes the latency, and throughput, etc, and then
generates reports that can be consumed in Jenkins. We also use Jenkins to
visualize the trend of performance changes which is very useful to find
regressions.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
