PELS: Performance Engineering for (sparse) Linear Solvers Demo

This is the hands-on material for PELS tutorial. To follow the tutorial, please open PELS.ipynb in Jupyter notebook and follow the tutorial there.

Terminal version

If you prefer doing experiments from teh command-line, you can execute the Python files kernels.py and pcg.py. The flag --help will describe how to run certain benchmarks and demos, e.g.:

python3 pcg.py --help

Notes for developing/updating

• When updating the Python programs, try to keep the test_*.py files up to date and make sure that nothing breaks by regularly running pytest.
• Before commiting changes to the Jupyter notebook PELS.ipynb, set up your local repo to filter out changes in execution counts etc.:

  pip install pre-commit
  pre-commit install