EuroHPC 2024 Modules¶

We present here the modules used to compile and run SIESTA under EuroHPC machines for the 2024 benchmarks. Be aware that these can be slightly different as time goes by, but it can provide a good starting point.

For the GPU partitions, you will also need a GPU-enabled version of ELPA. This can be already provided by the HPC center, or you can compile it yourself. You can find ELPA here: https://gitlab.mpcdf.mpg.de/elpa/elpa.git

Intel-based systems¶

Note that IntelMPI showed a few numerical issues and memory leaks, so we recommend using OpenMPI instead.

Leonardo DCGP¶

module load intel-oneapi-compilers/2023.2.1 intel-oneapi-mkl/2023.2.0 intel-oneapi-mpi/2021.10.0
module load netcdf-fortran/4.6.1--oneapi--2023.2.0 spack

spack load /g24wncv

Leonardo Booster (GPU)¶

module load gcc/12.2.0 fftw/3.3.10--openmpi--4.1.6--gcc--12.2.0
module load openblas/0.3.24--gcc--12.2.0 netlib-scalapack/2.2.0--openmpi--4.1.6--gcc--12.2.0
module load netcdf-fortran/4.6.1--openmpi--4.1.6--gcc--12.2.0 cuda/12.1
module load cmake ninja

Marenostrum 5 GPP¶

module load intel/2023.2.0 mkl/2023.2.0 ucx/1.15.0 bsc/1.0 ninja/1.11.1 cmake/3.25.1 openmpi/4.1.5
module load pnetcdf/1.12.3-openmpi hdf5/1.14.1-2-openmpi netcdf/c-4.9.2_fortran-4.6.1_cxx4-4.3.1_hdf5-1.14.1-2_pnetcdf-1.12.3-openmpi
module load fftw/3.3.10-openmpi

Marenostrum 5 ACC (GPU)¶

module load EB/apps libtool Autoconf Python
module load ScaLAPACK/2.2.0-gompi-2023b-fb FFTW/3.3.10-GCC-13.2.0 CUDA/12.5.0
module load netCDF-Fortran GCC/12.3.0
module load CMake/3.26.3-GCCcore-12.3.0 Ninja/1.11.1-GCCcore-12.3.0

Cray-based systems¶

LUMI-C¶

module load LUMI/24.09 partition/C libreadline PrgEnv-cray
module load cray-hdf5-parallel cray-parallel-netcdf cray-netcdf-hdf5parallel cray-fftw

LUMI-G (GPU)¶

module load LUMI/24.09 partition/G PrgEnv-cray
module load cray-hdf5-parallel cray-parallel-netcdf cray-netcdf-hdf5parallel
module load cray-fftw rocm buildtools

GNU-based systems¶

Deucalion¶

module load ScaLAPACK/2.2.0-gompi-2023a-fb netCDF-Fortran/4.6.1-gompi-2023a FFTW/3.3.10-GCC-12.3.0
module load CMake Ninja libreadline/8.2-GCCcore-12.3.0 Bison

Deucalion (GPU)¶

module load ScaLAPACK/2.2.0-gompi-2023a-fb netCDF-Fortran/4.6.1-gompi-2023a FFTW/3.3.10-GCC-12.3.0
module load CMake Ninja libreadline/8.2-GCCcore-12.3.0
module load Autoconf/2.71 libtool/2.4.7
module load CUDA/12.4.0

Discoverer¶

module load gcc/latest openblas/0/latest-gcc lapack/latest-gcc scalapack/2/latest-gcc-openmpi ucx/1/latest cmake/latest
module load libfabric/1/latest libaec/1/latest-gcc openmpi/4/gcc/latest zstd/1/latest-gcc zlib/1/latest-gcc fftw/3/latest-gcc-openmpi
module load hdf5/1/1.14/latest-gcc-openmpi netcdf/c/4.9/latest-gcc-openmpi pnetcdf/1/latest-gcc-openmpi netcdf/fortran/4.6/latest-gcc-openmpi
export OMP_NUM_THREADS=1 # Required for openblas detection.

Karolina¶

module load GCC CMake OpenBLAS LAPACK ScaLAPACK/2.2.0-gompi-2023b-fb libreadline Bison/3.8.2-GCCcore-13.2.0
module load PnetCDF/1.12.3-gompi-2022a netCDF-Fortran/4.6.1-gompi-2023b FFTW.MPI/3.3.10-gompi-2023b

Karolina (GPU)¶

module load GCC/12.3.0 CMake/3.26.3-GCCcore-12.3.0 OpenBLAS LAPACK ScaLAPACK/2.2.0-gompi-2023b-fb
module load libreadline CUDA/12.3.0 Autoconf libtool
module load PnetCDF/1.12.3-gompi-2022a netCDF-Fortran/4.6.1-gompi-2023b FFTW.MPI/3.3.10-gompi-2023b

Meluxina¶

module load ScaLAPACK/2.2.0-gompi-2023a-fb netCDF-Fortran/4.6.1-gompi-2023a FFTW/3.3.10-GCC-12.3.0
module load CMake Ninja libreadline/8.2-GCCcore-12.3.0 Bison

Meluxina (GPU)¶

module load ScaLAPACK/2.2.0-gompi-2023a-fb netCDF-Fortran/4.6.1-gompi-2023a FFTW/3.3.10-GCC-12.3.0
module load CMake Ninja libreadline/8.2-GCCcore-12.3.0
module load Autoconf/2.71 libtool/2.4.7 Automake foss Perl
module load CUDA/12.2.0

Vega¶

module load netCDF-Fortran PnetCDF/1.12.3-gompi-2021b libreadline
module load OpenBLAS ScaLAPACK FFTW/3.3.10-GCC-12.3.0
module load CMake Ninja Bison/3.8.2-GCCcore-12.3.0

Vega (GPU)¶

module load netCDF-Fortran PnetCDF/1.12.3-gompi-2021b libreadline
module load OpenBLAS ScaLAPACK FFTW/3.3.10-GCC-12.3.0
module load Autoconf libtool
module load GCC CUDA
module load CMake Ninja