[molpro-user] how to run molpro across an opteron cluster

[Dr Seth OLSEN]

Hi Molpro Users,

I have built parallel Molpro to run on an 64bit Opteron node running Fedora Core 2.    Each node in the cluster has 2 cpus, each of which represents one  slot in the context of the Sun Grid Engine queuing system which manages the job load on the cluster.  We are currently working on creating a parallel environment which will allow molpro jobs to use both slots on a single node.  Is there anyone on the list with experience in similar matters?  After this, we will work on getting molpro jobs to run across nodes.  Regarding this, I have some questions about how parallel molpro does machine-machine communications.  How does Molpro do internode communication?  Right now, the nodes of the cluster are set up to speak with each other via ssh.  What must be done to get molpro to work across nodes in that environment?

Cheers,

Seth Olsen


ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms

Dr Seth Olsen, PhD
Postdoctoral Fellow, Computational Systems Biology Group
Centre for Computational Molecular Science
Chemistry Building,
The University of Queensland
Qld 4072, Brisbane, Australia

tel (617) 33653732
fax (617) 33654623
email: s.olsen1 at uq.edu.au
Web: www.ccms.uq.edu.au 

ccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccmsccms