Configure support tools

UniversityOfGdanskTeamPython / epseon_backend

Vulkan based GPU accelerated implementation of numerical Schrödinger equation solver

MIT License

0 stars 0 forks source link

Configure support tools #5

Closed Argmaster closed 11 months ago

Argmaster commented 1 year ago

(Mostly) Python:

pre-commit
ruff
poetry
pytest
pytest-cov
codecov?
prettier
mkdocs

https://github.com/Argmaster/pygerber https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks https://python-poetry.org/docs/configuration/ https://docs.astral.sh/ruff/configuration/ https://pypi.org/classifiers/ https://semver.org/ https://docs.github.com/en/get-started/quickstart/github-flow https://docs.github.com/en/actions

Argmaster commented 11 months ago

I have configured C++ as Python extension builds and CI for it. Additionally I have identified major problem with our approach to testing - we are creating GPU compute algorithm and we intended to use free Github Actions devices wich don't have GPUs :D I'm currently trying to create some workaround for it by using siftshader - Vulkan implementation for CPU.

Argmaster commented 11 months ago

I had problems with getting shiftshader to work correctly. It took me some research on ruinning Vulkan applications without a display and multiple reattempts (27 to be precise) but I have managed to resolve it. It was caused by shiftshader shared objects being incorrectly copied via SCP. Afterwards I have managed to fully configure Python & CMake C++ builds with Vulkan API, pybind11, GoogleTest and other tools. Now they are successfully running on GitHub-provided machines.