When you have python 3.9 to 3.14 (CPython, 64-bit, x86_64) installed,
python -m pip install pyoomph
should install the basic framework. On Mac silicon processor systems (arm64, e.g. M1-M4), please execute it in a Rosetta2 terminal (see below). For the maximum performance and system-specific information, please refer to the sections below.
If you cannot manage to install it, refer to our tutorial. If this cannot help, you can ask for help c.diddens@utwente.nl
For maximum performance, also install Microsoft Build Tools, available for download here.
Verify whether everything runs fine by
python -m pyoomph check all
If you encounter segmentation faults during solving, please try to downgrade your MKL package, e.g. via python -m pip install mkl==2024.1.0.
If you have installed via pip (see above), just make sure that you have the gcc compiler installed and check via
python -m pyoomph check all
If you encounter segmentation faults during solving, please try to downgrade your MKL package, e.g. via python -m pip install mkl==2024.1.0.
If you have a recent Mac with an Apple silicon (arm64) chip, you must run all commands in a Rosetta 2 terminal, see here how to set it up (note: recent systems must be handled differently, see here). Also, please downgrade mkl by
python3 -m pip install mkl==2021.4.0
Since the development of Intel MKL (containing the fast MKL Pardiso solver) was discontinued for Mac, using this solver might require installing a less recent version of python. See Mac_arm64_with_Pardiso.md for more details.
Make sure to have the XCode developer tools, e.g. by installing them via
xcode-select --install
and test pyoomph via
python -m pyoomph check all
Alternatively, if you do not want to use the Rosetta 2 terminal detour, you can also install it directly on arm64 systems. Vatsal Sanjay from CoMPhy Lab has contributed an installation script
bash installOnArm.sh
to perform this installation. The only downside is that the fast MKL Pardiso solver is not supported on native arm64 systems. For optimal performance, in particular for larger problems, one therefore should consider the Rosetta 2 terminal way for the time being.
First, you have to make sure to have installed all dependencies, including some development files (headers). On e.g. Ubuntu, you can do the following
sudo apt-get install libopenmpi-dev pybind11-dev libginac-dev libcln-dev libgmp-dev ccache libmkl-rt
On other Linux distributions, other package manager like yum or pacman can be used to install the same libraries and headers.
Afterwards, compile the required parts of oomph-lib by running
bash ./prebuild.sh
in your pyoomph directory.
Install required and optional python modules via
python -m pip gmsh mkl mpi4py matplotlib numpy petsc4py pybind11 pygmsh scipy meshio pybind11-stubgen setuptools
If you want to install pyoomph for development, it is best to install it via
bash ./build_for_develop.sh
In that case, changes you make in the python files of pyoomph will be used automatically. Alternatively, for a system- or user-wide installation by installing pyoomph, do
bash ./install.sh
as last step instead.
Verify whether everything runs fine by
python -m pyoomph check all
Besides XCode, you must install a few third-party tools. This can be done by e.g. Homebrew:
brew install openmpi ccache ginac
Since the development of Intel MKL (containing the fast MKL Pardiso solver) was discontinued for Mac, using this solver might require installing a less recent version of python. See Mac_arm64_with_Pardiso.md for more details.
Restart your (Rosatta) terminal afterwards. Install required python modules via
python3 -m pip install pybind11 gmsh commonmark six pyparsing pygments pillow numpy mpi4py kiwisolver fonttools cycler scipy rich python-dateutils packaging meshio matplotlib pygmsh pybind11-stubgen
If you have a recent Mac with an Apple silicon (arm64) chip, install
python3 -m pip install mkl==2021.4.0
otherwise, you can install
python3 -m pip install mkl
Then, compile the required parts of oomph-lib:
bash ./prebuild.sh
in your pyoomph directory.
If you want to install pyoomph for development, it is best to install it via
bash ./build_for_develop.sh
In that case, changes you make in the python files of pyoomph will be used automatically. Alternatively, for a system- or user-wide installation by installing pyoomph, do
bash ./install.sh
as last step instead.
Verify whether everything runs fine by
python -m pyoomph check all