I am trying to use Abaqus/Explicit on a research cluster. Our group has access to 4 nodes, with 24 cpus on each node. I created a SLURM file, which would have used 64 cpus(tasks) spread across the 4 nodes with 16 tasks per node.

However, I received an error message saying that only 24 cpus were available and to modify my Mp_host_file in  abaqus_v6.env. I couldn't find out how to do this exactly, but from what i read, it seems that this is to parallelize across multiple machines. I then decided to try with just the 24 cpus available, but it aborted without any error message.

Below is the last part of the submit.sh script where I try to execute the job.

## Run commandmodule

load openmpi

/path/to/abaqus job=filename.inp cpus=24 double=both mp_mode=mpi

I couldn't find much documentation on how to use abaqus across multiple cores, so any help would be appreciated.

Thanks