Hello,
On our cluster I get the message when I try to run on 32 cores on two nodes (each node has 16 cores):
HP-MPI licensed for Accelrys.
Host 0 -- ip 192.168.10.8 -- ranks 0 - 31
host | 0
======|======
0 : SHM
Prot - All Intra-node communication is: SHM
torque.sh is:
#!/bin/sh -f
# This file is submitted to the queuing system
# The command for the job submission:
# qsub -V -N MS_9MPIA -l nodes=2:ppn=16 -q default torque.sh
# Appropriate environment should be set before the job submission
...
and machines.LINUX.txt contain two nodes:
cn-mpi07
cn-mpi06
The job is running on only one node named Host 0.
I suppose that's because of HP-MPI not properly set up.
Can anyone give me some more clues ?
Thank you,
Calin Gabriel Floare