/minidolfin

Minimal FE library

Primary LanguagePythonGNU Lesser General Public License v3.0LGPL-3.0

minidolfin

minidolfin is experimental minimal FE library written in Python and using JIT-compilation techniques of FEniCS stack, TSFC, Numba.

Currently it uses PETSc as a linear algebra backend which is maybe a bit too heavy weight considering that minidolfin does not run with MPI parallelization (yet).

Getting started

To install in FEniCS dev-env Docker image:

git clone https://github.com/blechta/minidolfin.git
cd minidolfin
fenicsproject run quay.io/fenicsproject/dev-env

and in the container:

pip3 install --user -r requirements.txt
pip3 install --user -e .

To run a demo:

cd demo
python3 helmholtz.py -h
python3 helmholtz.py [args]

Authors

License

GNU LGPL v3 or any later version: COPYING, COPYING.LESSER.

Testing

https://circleci.com/gh/blechta/minidolfin.svg?style=svg