Re: [AMBER] Proper Algorithm for GPU?

From: case <case.biomaps.rutgers.edu>
Date: Sun, 15 Dec 2013 13:06:26 -0500

On Sat, Dec 14, 2013, X Sun wrote:
>
> I am using PMEMD.CUDA on GPU, it works very well. My advisor asked me to
> write a review to collect proper algorithm for GPU and list the reason
> why the algorithm is suitable for the parallel computing on GPU. It's
> easy to find some applications such as AMBER on GPU. But how can I
> explain certain application is suitable for parallel computing? Could
> you give me some suggestion?

It's not clear whether your reading has included the three papers listed
under "Authorship and Support" on this page:

   http://ambermd.org/gpus/

Those papers give a lot of details about how Amber was adapted for GPUs.

(I'm not sure if this is what you are asking for or not; finding all
"algorithms" that perform well on GPU's would be a very different task.)

...dac


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Dec 15 2013 - 10:30:02 PST
Custom Search