At my university we have 50 computers in the lab. We would like to use a scheduler to schedule our fluid models, and I was wondering what is a good suggestion?
TIA
On 09/10/08 18:22, Mag Gam wrote:
At my university we have 50 computers in the lab. We would like to use a scheduler to schedule our fluid models, and I was wondering what is a good suggestion?
SGE (Sun Grid Engine):
http://gridengine.sunsource.net/ http://www.sun.com/software/gridware/ http://en.wikipedia.org/wiki/Sun_Grid_Engine
Or if you've got a lot of money you need to burn, you can use LSF:
http://www.platform.com/products/platform-lsf
Mag Gam wrote:
At my university we have 50 computers in the lab. We would like to use a scheduler to schedule our fluid models, and I was wondering what is a good suggestion?
you might look at using one of the scientific clustering packages, like Oscar, which implements and manages an MPI cluster. this of course assumes your fluid model software is written to use MPI
On Wed, 10 Sep 2008 19:54:44 -0700 "JRP" == John R Pierce <John> wrote:
JRP> Mag Gam wrote: >> At my university we have 50 computers in the lab. We would like >> to use a scheduler to schedule our fluid models, and I was >> wondering what is a good suggestion? >>
JRP> you might look at using one of the scientific clustering JRP> packages, like Oscar, which implements and manages an MPI JRP> cluster. this of course assumes your fluid model software is JRP> written to use MPI
If you're going for cluster software you might consider
It's even based on CentOS.
Here at our place we have a dedicated cluster with Rocks. In addition I took the SGE-rpm from the distribution and installed it on our regular (CentOS) workstations so that jobs can be scheduled on these machines too
Thankyou everyone. How does SGE compare? Is it easy to implement? What about its features compared to others? How is the code quality, stability, and documentation?
TIA
On Thu, Sep 11, 2008 at 4:15 AM, Bernhard Gschaider bgschaid_lists@ice-sf.at wrote:
On Wed, 10 Sep 2008 19:54:44 -0700 "JRP" == John R Pierce <John> wrote:
JRP> Mag Gam wrote:
At my university we have 50 computers in the lab. We would like to use a scheduler to schedule our fluid models, and I was wondering what is a good suggestion?
JRP> you might look at using one of the scientific clustering JRP> packages, like Oscar, which implements and manages an MPI JRP> cluster. this of course assumes your fluid model software is JRP> written to use MPI
If you're going for cluster software you might consider
It's even based on CentOS.
Here at our place we have a dedicated cluster with Rocks. In addition I took the SGE-rpm from the distribution and installed it on our regular (CentOS) workstations so that jobs can be scheduled on these machines too _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos