Skip to content

Latest commit

 

History

History
23 lines (12 loc) · 1.33 KB

README.md

File metadata and controls

23 lines (12 loc) · 1.33 KB

Periodic profiling of PorePy

This repository contains the runscripts for the periodic profiling of PorePy, which allows to trace performance improvement or regression among different commits of the develop branch. The runscripts are based on the asv package.

The report can be found here: https://pmgbergen.github.io/porepy-profiling/

Write your benchmarks

The benchmark cases must be located in the benchmarks/ folder. Commiting them into the repository will do the job and they will appear in the report when the periodic job runs, typically once a day. To write your benchmark case, see the asv tutorial.

Before pushing the benchmark case, test if it works correctly:

asv run --python=same --quick --dry-run --launch-method=spawn --show-stderr

The actual run command (if you want to run the full benchmark suite locally):

asv run --launch-method=spawn --show-stderr

Other useful commands: asv publish generates html reports, asv preview opens the report in a browser.

Manual profiling

To investigate your program performance, we suggest the viztracer package. See the quickstart runscript for it: run_viztracer.py.