On Fri, Apr 22, 2011 at 9:41 AM, Kyle Klein
<kylet...@gmail.com> wrote:
Hi Matt,
The assignment says to try running on multiple nodes, up to 32. Do we have to do anything special to get them? I've tried what triton resources say where you set nodes=1:ppn=32 and the queue to large, however the speedup beyond 4 cores is extremely limited.
I am not sure if this is because each processor has only 4 cores, or because the code is not actually running on that many processors.
Am I missing something?
best,
Kyle
--
Kyle Klein
Ph.D. Student
Department of Computer Science UC Santa Barbara