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

From: Gould, Ian R <i.gould.imperial.ac.uk>
Date: Fri, 21 Sep 2012 08:04:58 +0000

Dear Tru,

You will have to check very carefully the spec of the motherboard, most
will only run at x8 PCI-E when you have 3 or more graphics cards in them
and that will cripple even non mpi performance.

Also you will need a huge power supply, a huge case and lots of
cooling.... I would suggest building two machines each with two 690's in
them, it will probably work out the same cost or only a little more
expensive. Same applies if you want to put 4 680's in a single case.

Cheers
Ian

Women love us for our defects. If we have enough of them, they will
forgive us everything, even our intellects.
Oscar Wilde, 
-- 
Dr Ian R Gould
Reader in Computational Chemical Biology
Department of Chemistry
Imperial College London
Exhibition Road
London
SW7 2AY
E-mail i.gould.imperial.ac.uk
http://www3.imperial.ac.uk/people/i.gould
Tel +44 (0)207 594 5809
On 21/09/2012 07:54, "Tru Huynh" <tru.pasteur.fr> wrote:
>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
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 21 2012 - 01:30:02 PDT
Custom Search