[gmx-users] Parallel GPU calculation on K80

Szilárd Páll pall.szilard at gmail.com
Fri Apr 22 14:36:06 CEST 2016

-gpu_id provides the *per node* GPU to PP rank mapping, so you don't
specify all GPUs, but list them per node in the order you want to assign
them to your MPI ranks.


On Fri, Apr 22, 2016 at 10:17 AM, Yan Wang <wangyanforward at gmail.com> wrote:

> Dear Gromacs Users,
> Recently I have several calculation jobs which run on the K80 using GPU
> acceleration. There was six K80 cards in my cluster. As you all know, each
> K80 card has two cores. There were 12 number of the GPU in my cluster. When
> I used the option -gpu_id 01234567891011, the software did not recognized
> the last two cores of the K80 card. So is there any suggestions to solve
> this problem?
> Thank you!
> ===================================
> *Dr. Yan WANG*
> Department of Biology
> South University of Science and Technology of China
> No.1088 Xueyuan Blvd, Nanshan District, Shenzhen, P.R. China 518055
> Tel.+86-182-4497-5993
> Email: wangyanforward at gmail.com
> ====================================
> --
> Gromacs Users mailing list
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list