Hm.. I think it is likely to to be Moab not scheduling those jobs fast enough.

I've opened a support ticket to AdaptiveComputing to try to get it
updated to the latest version.

You can open another shell and qrun the job, if it runs quickly then
it's Moab not doing its job as fast as we would like.

Robin


On Fri, Jan 10, 2014 at 11:27 AM, Mueller, Jens Dr. <[log in to unmask]> wrote:
> This is much slower than what I’ve observed before the upgrade. I takes more
> than a 1 minute for me just to get a single processor through “qsub -IV -l
> nodes=1:ppn=1 -l walltime=80:00:00”. This might hint to an issue with Torque
> getting the jobs launched in general.
>
> Jens