PSYDAC is a Python 3 Library for isogeometric analysis.
- Requirements
- Python setup and project download
- Installing the library
- Optional PETSc installation
- Uninstall
- Running tests
- Speeding up Psydac's core
- User Documentation
- Code Documentation
Psydac requires a certain number of components to be installed on the machine:
- Fortran and C compilers with OpenMP support
- OpenMP library
- BLAS and LAPACK libraries
- MPI library
- HDF5 library with MPI support
The installations instructions depend on the operating system and on the packaging manager used.
To install all requirements on a Linux Ubuntu operating system, just use APT, the Advanced Packaging Tool:
sudo apt update
sudo apt install python3 python3-dev python3-pip
sudo apt install gcc gfortran
sudo apt install libblas-dev liblapack-dev
sudo apt install libopenmpi-dev openmpi-bin
sudo apt install libomp-dev libomp5
sudo apt install libhdf5-openmpi-dev
To install all the requirements on a macOS operating system we recommend using Homebrew:
brew update
brew install gcc
brew install openblas
brew install lapack
brew install open-mpi
brew install libomp
brew install hdf5-mpi
Please see the instructions for the pyccel library for further details.
We recommend creating a clean Python virtual environment using venv:
python3 -m venv <ENV-PATH>
where <ENV-PATH>
is the location to create the virtual environment.
(A new directory will be created at the required location.)
In order to activate the environment from a new terminal session just run the command
source <ENV-PATH>/bin/activate
One can clone the Psydac repository at any location <ROOT-PATH>
in the filesystem which does not require administrator privileges, using either
git clone https://github.com/pyccel/psydac.git
or
git clone git@github.com:pyccel/psydac.git
The latter command requires a GitHub account.
Psydac depends on several Python packages, which should be installed in the newly created virtual environment.
These dependencies can be installed from the cloned directory <ROOT-PATH>/psydac
with the following steps.
First, set an environment variable with the path to the parallel HDF5 library. This path can be obtained with a command which depends on your system.
-
Ubuntu/Debian:
export HDF5_DIR=$(dpkg -L libhdf5-openmpi-dev | grep "libhdf5.so" | xargs dirname)
-
macOS:
export HDF5_DIR=$(brew list hdf5-mpi | grep "libhdf5.dylib" | xargs dirname | xargs dirname)
Next, install the Python dependencies using pip
:
export CC="mpicc"
export HDF5_MPI="ON"
python3 -m pip install --upgrade pip
python3 -m pip install -r requirements.txt
python3 -m pip install -r requirements_extra.txt --no-build-isolation
At this point the Psydac library may be installed in standard mode, which copies the relevant files to the correct locations of the virtual environment, or in development mode, which only installs symbolic links to the Psydac directory. The latter mode allows one to effect the behavior of Psydac by modifying the source files.
-
Standard mode:
python3 -m pip install .
-
Development mode:
python3 -m pip install --editable .
Although Psydac provides several iterative linear solvers which work with our native matrices and vectors, it is often useful to access a dedicated library like PETSc. To this end, our matrices and vectors have the method topetsc()
, which converts them to the corresponding petsc4py
objects.
(petsc4py
is a Python package which provides Python bindings to PETSc.) After solving the linear system with a PETSc solver, the function petsc_to_psydac
allows converting the solution vector back to the Psydac format.
In order to use these additional feature, PETSc and petsc4py must be installed as follows. First, we download the latest release of PETSc from its official Git repository:
git clone --depth 1 --branch v3.21.4 https://gitlab.com/petsc/petsc.git
Next, we specify a configuration for complex numbers, and install PETSc in a local directory:
cd petsc
export PETSC_DIR=$(pwd)
export PETSC_ARCH=petsc-cmplx
./configure --with-scalar-type=complex --with-fortran-bindings=0 --have-numpy=1
make all check
cd -
Finally, we install the Python package petsc4py
which is included in the PETSc
source distribution:
python3 -m pip install wheel Cython numpy
python3 -m pip install petsc/src/binding/petsc4py
- Whichever the install mode:
python3 -m pip uninstall psydac
- If PETSc was installed:
python3 -m pip uninstall petsc4py
The non-Python dependencies can be uninstalled manually using the package manager. In the case of PETSc, it is sufficient to remove the cloned source directory given that the installation has been performed locally.
Let <PSYDAC-PATH>
be the installation directory of Psydac.
In order to run all serial and parallel tests which do not use PETSc, just type:
export PSYDAC_MESH_DIR=<PSYDAC-PATH>/mesh/
python3 -m pytest --pyargs psydac -m "not parallel and not petsc"
python3 <PSYDAC-PATH>/mpi_tester.py --pyargs psydac -m "parallel and not petsc"
If PETSc and petsc4py were installed, some additional tests can be run:
python3 -m pytest --pyargs psydac -m "not parallel and petsc"
python3 <PSYDAC-PATH>/mpi_tester.py --pyargs psydac -m "parallel and petsc"
Many of Psydac's low-level Python functions can be translated to a compiled language using the Pyccel transpiler. Currently, all of those functions are collected in modules which follow the name pattern [module]_kernels.py
.
The classical installation translates all kernel files to Fortran without user intervention. This does not happen in the case of an editable install, but the command psydac-accelerate
is made available to the user instead. This command applies Pyccel to all the kernel files in the source directory, and the C language may be selected instead of Fortran (which is the default).
- Only in development mode:
python3 /path/to/psydac/psydac_accelerate.py [--language LANGUAGE] [--openmp]
Find our latest code documentation here.
After installation, a command psydac-mesh
will be available.
psydac-mesh -n='16,16' -d='3,3' square mesh.h5