Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Bly, MJ (Martin) said:
> That which defines a 'CPU' varies - some say it refers to a
> chip and others to a core. 'CPU' count varies between nodes:
> for example all our boxes have two chips but some have dual
> cores and I expect we may have quad cores to deal with. (And
> don't get me started on hyperthreading chips...). Is SMPSize
> supposed to be defined as n-chips or n-cores?
> How does oteh schema deal with single, dual and quad chip
> boxes in the same batch system?
SMPSize was defined long before multi-core CPUs were a twinkle in
Intel's eye, but in practice I think it should clearly be the number of
cores as they are treated as separate CPUs for most purposes. The fact
that our current way of working doesn't account for heterogeneous sets
of WNs is a different can of worms, also being considered in Steve's WG.
In practice you currently either have to define different batch queues
for different classes of WN, or publish a minimum/typical spec and hope
for the best.
Stephen
|