Do you like seeing unending warning messages about Volta compatibility? If
so then hurry to install CUDA 9!
Otherwise, I don't think there are any features in that tool kit that do
any good for Amber whatsoever. Anyone can chime in here if I'm missing
something.
On the other hand if you are doing deep learning there are significant
speedups for small batch sizes and convolutions, just nothing for molecular
dynamics.
On Oct 31, 2017 10:54 AM, "Ross Walker" <ross.rosswalker.co.uk> wrote:
> Hi Robert,
>
> LOL, one has to love NVIDIA's marketing folk. If only life was that easy...
>
> CUDA 9 isn't supported yet by AMBER 16 and probably won't be supported
> until AMBER 18 so my advice would be to stick with CUDA 8 for now.
>
> All the best
> Ross
>
> > On Oct 31, 2017, at 1:11 PM, Weinzierl, Robert O J <
> r.weinzierl.imperial.ac.uk> wrote:
> >
> > I was wondering whether installing the latest CUDA release (CUDA 9;
> https://developer.nvidia.com/cuda-toolkit/whatsnew) has any beneficial
> effects on running Amber pmemd.cuda.
> >
> > Nivida claims that "with CUDA 9 you can speed up your applications while
> making them more scalable and robust.". Has anybody tried this out yet? Or
> shall I stick with CUDA 8 for the time being?
> >
> > Dr. Robert O.J. Weinzierl,
> >
> > Imperial College London,
> > Department of Life Sciences,
> > Sir Alexander Fleming Building, Room 510
> >
> > London SW7 2AZ; United Kingdom
> > _______________________________________________
> > 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Oct 31 2017 - 13:00:02 PDT
