The only compulsory pre-requisites needed are compilers for C/C++/Fortran (such as the GNU compiler collection) and a Python interpreter.
In addition, several optional libraries can be used by code_saturne for a more complete feature set.
These links are provided for information. Most of these libraries are available directly in compiled form through the package managers of the Linux distributions. It is strongly advised to use these compiled packages when available.
On most Linux distributions, packages are separated in a <package> (needed to run) and <package>-dev (development, needed to compile) scheme, so do not forget to check that the development headers for required packages are also installed.
The installation manual provides more details on optional libraries and versions. Only the most common prerequisites are listed here.
Graphical interface
- PyQt 5
Mesh formats for pre- and post-processing
Parallel computing
- An MPI implementation, e.g. OpenMPI or MPICH
- PT-Scotch (optional, but recommended for a better domain partitioning on compute clusters)
Conjugate heat transfer
code_saturne has some built-in conjugate heat transfer functionality, but for cases where some of its additional features are needed, coupling with the Syrthes code may be used.
As Syrthes does not have a dedicated web-site we provide an updated and slightly modified Syrthes 5.0 archive. Compared to the older reference distribution from the EDF web site, this archive is patched to fix compilation on newer systems, and does not include older HDF5, MED, Metis and Scotch distributions, as those may already be handled by the code_saturne installer or trough system packages. installing the GUI requires PyQt5 (usually available as a package) and optionally PyQtChart, which can be installed using pip if not available through a Linux distribution package. An older Syrthes 4.3.5 archive is also available.