Re: [Condor-users] Fluent + Condor

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Dec 20, 2011, at 9:55 AM, Tiago Macarios wrote:

> I want to run FLUENT job using condor. I found the submit file on a thread from 2009:
> 
> universe  =  vanilla
> requirements =
> getenv = true
> executable      = /opt/Fluent.Inc/bin/fluent
> arguments = 3d -g -t<nprocs> -cnf=host.txt
> input = /cfd/gunaydin/3d/journal_file
> output          = output_file
> queue
> 
> But I dont know how to generate the hosts.txt file. In other words,
> How can I tell fluent which machines condor "reserved" for the job? How can get the executing machine before running the fluent command line?

With this submit file, your job will be given one machine to run one. One core on one machine, to be exact. For parallel applications, you need to use Condor's parallel universe. Section 2.9 of the Condor Manual has details:
http://research.cs.wisc.edu/condor/manual/v7.6/2_9Parallel_Applications.html

Thanks and regards,
Jaime Frey
UW-Madison Condor Team