FluidSim is part of the FluidDyn project. Some issues regarding the installation of Python packages are discussed in the main documentation of the project.
Python 2.7 or >= 3.4
fluidsim needs fluidfft. If you don’t install it before carefully, it will be installed automatically and you won’t be able to use fancy FFT libraries (using for example MPI with 2D decomposition or GPU with CUDA). If you are not too concerned about performance, no problem. Otherwise, install fluidfft as explained here.
A C++11 compiler (for example GCC 4.9 or clang)
We choose to use the new static Python compiler Pythran for some functions. Our microbenchmarks show that the performances are as good as what we are able to get with Fortran or C++!
To reach good performance, we advice to try to put in the file ~/.pythranrc the lines (it seems to work well on Linux, see the Pythran documentation):
[pythran] complex_hook = True
The compilation of C++ files produced by Pythran can be long and can consume a lot of memory. If you encounter any problems, you can try to use clang (for example with
conda install clangdev) and to enable its use in the file ~/.pythranrc with:
[compiler] CXX=clang++ CC=clang
h5py (optionally, with MPI support, but only if you know what you do)
Prebuilt installations (for eg. via h5py wheels) may lack MPI support. Most of the time, this is what you want. However, you can install h5py from source and link it to a hdf5 built with MPI support, as follows:
$ CC="mpicc" HDF5_MPI="ON" HDF5_DIR=/path/to/parallel-hdf5 pip install --no-deps --no-binary=h5py h5py $ python -c 'import h5py; h5py.run_tests()'
See the h5py documentation for more details.
Optionally (for MPI runs), mpi4py (which depends on a MPI implementation).
Basic installation with pip¶
If you are in a hurry and that you are not really concerned about performance, you can use pip:
pip install fluidsim
pip install fluidsim --user
You can also configure the installation of fluidsim by creating the file
~/.fluidsim-site.cfg and modify it to fit your requirements before the
installation with pip:
wget https://bitbucket.org/fluiddyn/fluidsim/raw/default/site.cfg.default -O ~/.fluidsim-site.cfg
Install from the repository (recommended)¶
Get the source code¶
For fluidsim, we use the revision control software Mercurial and the main repository is hosted here in Bitbucket. Download the source with something like:
hg clone https://bitbucket.org/fluiddyn/fluidsim
If you are new with Mercurial and Bitbucket, you can also read this short tutorial.
For particular installation setup, copy the default configuration file:
cp site.cfg.default site.cfg
and modify it to fit your requirements.
If you care about performance, correctly set up a configuration file. By default, some Pythran files are not Pythranized!
Build/install in development mode (with a virtualenv or with conda):
python setup.py develop
or (without virtualenv):
python setup.py develop --user
Of course you can also install FluidDyn with the install command
Run the tests!¶
You can run some unit tests by running
make tests (shortcut for
fluidsim-test -v) or
make tests_mpi (shortcut for
mpirun -np 2
fluidsim-test -v). Alternatively, you can also run
python -m unittest
discover from the root directory or from any of the “test” directories.