For most Scala projects, when I submit a Pull Request and I have to evaluate the performance impact of a Pull Request, I often have to manually do the following process:
- Checkout each distinct version, with and without the changes.
- Run the
sbt:jmh command, a same command, in each of them.
- Retrieve and store the results from each run.
- Arrange the results of both runs onto a spreadsheet, which I then use to show the relative error of each measure, and the observed improvement.
It may be desirable if sbt:jmh plugin could offer a command to carry ouy these operations directly, and were able to generate a single summary table, which included the score and error for each commit, and the relative difference between them.
Benchmark Mode Cnt Score(master) Error Score(branch) Error Change Units
Fili thrpt 10 45000 42 42000 41 93.3 % B/s
For most Scala projects, when I submit a Pull Request and I have to evaluate the performance impact of a Pull Request, I often have to manually do the following process:
sbt:jmhcommand, a same command, in each of them.It may be desirable if
sbt:jmhplugin could offer a command to carry ouy these operations directly, and were able to generate a single summary table, which included the score and error for each commit, and the relative difference between them.