Testbed Support for GridPP member institutes [mailto:[log in to unmask]] On Behalf Of Sam Skipsey said:
> (The graph from the slide doesn't help, as it implies that ATLAS MC is fine with an average of 1GB / core of physical memory.)
It may be worth pointing out that atlas MC jobs are not by any means all equal. In particular, higher pileup, as we're likely to see this year, is likely to mean more memory needed. (It used to be that very high pileup jobs essentially couldn't run at all, but I think it's been improved.)
Stephen
|