S3COM can be executed localy as well as on supercomputers. Note that all libraries, including RTTOV, must be compiled with openMP or similar to allow for parallelization (see rttov_nthreads option in namelist).
Levante
Example of modules to load to run RTTOV with multiple threads.
With intel compilers:
module purge
module load intel-oneapi-compilers/2022.0.1-gcc-11.2.0 \
openmpi/4.1.2-intel-2021.5.0 \
netcdf-fortran/4.5.3-openmpi-4.1.2-intel-2021.5.0 \
netcdf-c/4.8.1-openmpi-4.1.2-intel-2021.5.0 \
hdf5/1.12.1-openmpi-4.1.2-intel-2021.5.0 \
eccodes/2.21.0-intel-2021.5.0
With gcc compilers:
module load gcc/11.2.0-gcc-11.2.0 \
openmpi/4.1.2-gcc-11.2.0 \
parallel-netcdf/1.12.2-openmpi-4.1.2-gcc-11.2.0 \
netcdf-fortran/4.5.3-openmpi-4.1.2-gcc-11.2.0 \
netcdf-c/4.8.1-openmpi-4.1.2-gcc-11.2.0 \
hdf5/1.12.1-openmpi-4.1.2-gcc-11.2.0 \
eccodes/2.21.0-gcc-11.2.0
Job script (edit project number):
#!/bin/bash
#SBATCH -J S3COM # Specify job name
#SBATCH -p shared # Use partition shared
#SBATCH --nodes=1 # Number of requested nodes
#SBATCH --ntasks-per-node=48 # Tasks per nodes
#SBATCH -A PROJECT_NUMBER # Charge resources on this project account
#SBATCH -o s3com.o%j # File name for standard and error output
# Environment settings to run a MPI parallel program
module purge
module load intel-oneapi-compilers
module load openmpi/4.1.2-intel-2021.5.0
module load hdf5/1.12.1-openmpi-4.1.2-intel-2021.5.0
module load netcdf-fortran/4.5.3-openmpi-4.1.2-intel-2021.5.0
module load netcdf-c/4.8.1-openmpi-4.1.2-intel-2021.5.0
module load netcdf-cxx/4.2-openmpi-4.1.2-intel-2021.5.0
module load parallel-netcdf/1.12.2-openmpi-4.1.2-intel-2021.5.0
module load libaec/1.0.5-intel-2021.5.0
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/sw/spack-levante/hdf5-1.12.1-tvymb5/lib
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/sw/spack-levante/netcdf-fortran-4.5.3-k6xq5g/lib
# Limit stacksize ... adjust to your programs need and core file size
ulimit -s unlimited
ulimit -c 0
# Execute OpenMP programs
srun ./s3com
