Re: [AMBER] GPU MPI failed

From: Albert <mailmd2011.gmail.com>
Date: Thu, 11 Aug 2016 15:20:37 +0200

Hello:

thanks a lot for suggestions.

I noticed that it is the problem of the first step for the minimization.

The rest steps are all right to use multiple GPU running under Amber.

thank you very much.

Albert




On 08/11/2016 02:15 PM, David A Case wrote:
> 1. Please look in the test.mdout file to see if errors are reported there;
> 2. Please check that you job runs OK on a single GPU, and in parallel
> using pmemd.MPI. Note that (generally) minimizations should be
> done on a CPU, since dynamic range limitations in the cuda code
> can cause problems if there are bad contacts in the starting structure.
> 3. Please confirm that the pmemd.cuda.MPI test cases are passing. (This will
> help in understanding if the problem is with your pmemd.cuda.MPI
> installation, or with the particular problem you are using.)
>
> ...dac


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Aug 11 2016 - 06:30:03 PDT
Custom Search