Re: [AMBER] performance of pmed.cuda.MPI

From: Tru Huynh <tru.pasteur.fr>
Date: Fri, 21 Sep 2012 08:54:54 +0200

On Fri, Sep 14, 2012 at 11:46:28AM -0700, Scott Le Grand wrote:
> MPI performance of GTX 690 is abysmal because the two GPUs share the same
> PCIEX adaptor.
>
> That will improve down the road somewhat.
>
> In the meantime, I think you'll be happy at the performance of two
> independent runs (one on each GPU): 98+% efficiency when I last checked...

If I understand you correctly, with a 4 PCI-E motherboard, and 4x GTX 690,
one should run 8 independant pmemd.cuda (non MPI) to get the maximum throughput.

The GTX-690 is seen as 2 nvidia devices that are adressed independantly?
In order to get a better pmemd.cuda.MPI scaling, that one needs to only target
one of the 2 GPUS on each PCI-E for each run? How does that behave for independant
pmemd.cuda.MPI simulations? Do the shared PCI-E become the bottleneck?
Bottom line, are multiple GTX-690 in the same server worth it? or should
one stay with the regular GTX-680?

Best regards,

Tru
-- 
Dr Tru Huynh          | http://www.pasteur.fr/recherche/unites/Binfs/
mailto:tru.pasteur.fr | tel/fax +33 1 45 68 87 37/19
Institut Pasteur, 25-28 rue du Docteur Roux, 75724 Paris CEDEX 15 France  
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 21 2012 - 00:00:02 PDT
Custom Search