Hi All, I am the administrator for 200 Windows computers in our computer labs. Because of opening hours and sessions the computers are idle a minimum of 8 hours a day during session and mostly idle outside session. I’d like to make better use
of these computers and am considering how I can run batch jobs on the machines. The jobs we run are mostly CFD and use Ansys. As such we can’t link them HT Condor modules and it looks like we won’t be able to take checkpoints of our jobs. Is this then the case that we can’t checkpoint our jobs? Has anybody found a way to run batch jobs on their idle computers that can be checkpointed if required? I’m wondering if I can add a Linux VM to each of my machines and run the Condor on them. This way I can checkpoint the jobs if required.
I’d be grateful for any opinions particularly if you’ve tried this. Regards David
David Herd
UNSW AUSTRALIA |