>Ok, I make change, but still have sometime 2 nodes instead 4. > >If i run a simple script mpihello.pbs, my job run over the 4 nodes. > >However, I meet always trouble with our Ocean Model. Even if i keep same >PBS variable like test program Mpihello, my job run over 2 nodes instead 4. >By hand, lamboot, mpirun, lamhalt, all nodes compute.
I'm guessing you didn't implement the changes as recommended by Jeff Squyres, correct? His response is authoritative.
Oups, i look again what Jeff said and Lam Install guide.
I applied --with-pbs=/opt/pbs but the syntax is wrong : better is --with-boot-tm=/opt/pbs. Ok, i install again lam-7.0-ifort and will tell you about.
This result, i.e., "node3.cluster.ird.nc" is correct assuming this is the name of one of your compute nodes. Your job is started on one of the compute nodes, with the other nodes available to you (including the job-start node) listed in
PBS_NODEFILE=/var/spool/pbs/aux/174.editr.cluster.ird.nc
More about : When i look inside this directory, i never see this kind of file ? aux directory is empty. Why ?
According to the environment display, your job *originated* on
PBS_O_HOST=editr.cluster.ird.nc
that is, this is the node whence you ran qsub.
Many thanks to you
Best regard Jerome
------------------------------------------------------- This SF.net email is sponsored by: 2005 Windows Mobile Application Contest Submit applications for Windows Mobile(tm)-based Pocket PCs or Smartphones for the chance to win $25,000 and application distribution. Enter today at http://ads.osdn.com/?ad_id=6882&alloc_id=15148&op=click _______________________________________________ Oscar-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oscar-users
