[slurm-users] Multinode MPI job

Mahmood Naderan mahmood.nt at gmail.com
Fri Mar 29 14:54:32 UTC 2019


I found out that the standard script that specifies the number of tasks and
memory per cpu will do the same thing that I was expecting from packjob
(heterogeneous job).

#SBATCH --job-name=myQE
#SBATCH --output=big-mem
#SBATCH --ntasks=14
#SBATCH --mem-per-cpu=17G
#SBATCH --nodes=6
#SBATCH --partition=QUARTZ
#SBATCH --account=z5
mpirun pw.x -i mos2.rlx.in


What am I seeing then?
6 processes are on a node with 128GB memory (reserved 6*17=102GB)
1, 3, 2, 1 and 1 processes on 5 nodes each with 64GB memory.


Regards,
Mahmood
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20190329/b5a5f881/attachment.html>


More information about the slurm-users mailing list