Re: [mpich-discuss] mpiexec.hydra binding for multiple compute nodes
11 May
2015
11 May
'15
5:09 p.m.
Justin, I would also do #SBATCH -N 2 #SBATCH -n 32 since if you do #SBATCH -n 40 and then run mpiexec -np 32, Slurm will put 20 tasks on node 0 and 12 on node 1. With the #SBATCH -n 32, Slurm will give you a node with 16 or more cores, so, 20 core node will be used if core count <20 is not available.. MC -- Martin Cuma Center for High Performance Computing Department of Geology and Geophysics University of Utah _______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss
4090
Age (days ago)
4090
Last active (days ago)
0 comments
1 participants
participants (1)
-
Martin Cuma