Then I won't worry. Thanks for the responses!
:-) Lachele
Dr. B. Lachele Foley
Associate Research Scientist
Complex Carbohydrate Research Center
The University of Georgia
Athens, GA USA
lfoley.uga.edu
http://glycam.org
________________________________________
From: Scott Le Grand <varelse2005.gmail.com>
Sent: Wednesday, September 17, 2014 5:53 PM
To: AMBER Developers Mailing List
Subject: Re: [AMBER-Developers] GPU/cuda determinism
The SLA is same GPU/Toolkit...
There are cases where you *will* see deterministic trajectories beyond
this, but they are not the spec...
On Wed, Sep 17, 2014 at 12:49 PM, Niel Henriksen <shireham.gmail.com> wrote:
> Ross or Scott are probably the last word on this, but the
> answer is "no". My experience is that for a given model
> (say 780s), the results are always the same. But different
> models give different results likely because a different set
> of model-dependent rounding errors have occurred.
>
> --Niel
>
> On Wed, Sep 17, 2014 at 12:14 PM, B. Lachele Foley <lfoley.ccrc.uga.edu>
> wrote:
>
> > If I do two runs that are absolutely identical except one is run on a 780
> > and the other on a Titan Black, shouldn't I expect to get identical
> results?
> >
> >
> > :-) Lachele
> >
> > Dr. B. Lachele Foley
> > Associate Research Scientist
> > Complex Carbohydrate Research Center
> > The University of Georgia
> > Athens, GA USA
> > lfoley.uga.edu
> > http://glycam.org
> > _______________________________________________
> > AMBER-Developers mailing list
> > AMBER-Developers.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber-developers
> >
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Sep 17 2014 - 15:30:02 PDT
