GNU Queue - alternatives

Solution 1:

If you don't require a real queuing system, GNU parallel may suffice to start jobs on each system simultaneously. If you do need a real scheduler, then TORQUE Resource Manager and optionally a scheduler like Maui may be needed.

You might also be as well off with abandoning CentOS in favor of a live CD like PelicanHPC. At least then, the configuration would be simpler (for what it can do, at least). Assuming you're at a university of some sort, there's nobody there at all that does HPC and/or clustering? And no faculty with contacts at a national lab or similar facility that can offer CPU time to your project?

And this question may be a candidate to migrate to Serverfault.

Solution 2:

We use Condor for job queuing, etc.