Compiling and running FreeFem on a HPC

Dear FreeFem users,

I am trying to use FreeFem on a HPC cluster. The compilation with PETSc/SLEPc seems to be successful - FreeFem examples seem to work on a login node. However, when I try to run FreeFem using the scheduler SLURM, neither ff-mpirun, nor FreeFem alone (with, e.g., a simple array.edp example, without PETSc/SLEPc or paralellism) works - I get the following error:

/home/mvgopkm/FreeFem/FreeFem-sources/src/mpi/FreeFem++-mpi: error while loading shared libraries: libblas.so.3: cannot open shared object file: No such file or directory

So you have any idea how to circumvent this error? I am attaching a script with the compilation commands, the PETSc and FreeFem config.log files, and also the Slurm script. Any help is appreciated.
configure_petsc_complex.log (1.1 MB)
configure_petsc_real.log (6.3 MB)
ffCompilation.log (5.5 KB)
config_freefem.log (366.6 KB)

You need to add the path to your libblas.so in your LD_LIBRARY_PATH.

1 Like

Thanks, this indeed solved the problem!