Dear All,
In the past (LSF) it was possible to submit parallel jobs on several cores for ABAQUS. However, now with SLURM i encounter the problem stated below:
Abaqus Error: The number of cpus used is more than the total number of processors specified by the host list. mp_host_list: [['ncm0081.hpc.itc.rwth-aachen.de', 1]]. Abaqus/Analysis exited with error(s).
I suspect it is due to a fautly setting of the enviromental variable $R_WLM_ABAQUSHOSTLIST,
as the job submission script, found on the RWTH documentation website states:
cat << EOF > $abaqus_v6.env #verbose = 3 #ask_delete = OFF mp_file_system = (SHARED, LOCAL) mp_host_list = $R_WLM_ABAQUSHOSTLIST EOF
It seems like currently there is only one host in that list, which i suspect to be insufficient when running on, let say 16 cores.
Best Regards, Karthik
claix18-slurm-pilot@lists.rwth-aachen.de