On Thu, Sep 20, 2012 at 1:50 AM, marawan hussain
<marawanhussain.yahoo.com>wrote:
> Hi Jason,
> Thanks for the reply...What is weird for me is that the Cluster is a very
> new Tesla cluster and the connection of the infiniband type and the
> compilation was carried out using the new mvapich2 library. I'm aware of
> all this issues like NVT to NPT and PMEMD versus Sander..Although the
> previous cluster was old and the connection was ethernet, the scaling was
> much better....If you have some clues about the most potential possible
> source of errors in such cases i will appreciate your help...
>
Wait -- is this pmemd.cuda? pmemd.cuda doesn't scale past 8 GPUs since the
FFT takes ~1/8 of the computational time and can only be done on a single
GPU.
The point of my response was that you didn't provide enough information to
provide any more specific answer.
All the best,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 20 2012 - 05:00:02 PDT
