Hi Stephen,
I think what you're looking for is export CUDA_VISIBLE_DEVICES,
specifically something like-
export CUDA_VISIBLE_DEVICES=0
pmemd.cuda -O -i job1.in...
export CUDA_VISIBLE_DEVICES=1
pmemd.cuda -O -i job2.in...
Where 0 and 1 are the two cards-
http://ambermd.org/gpus/ has a lot more
details on running them under the section for multi GPUs.
Best,
Kenneth
On Tue, Dec 1, 2015 at 1:25 AM, Stephen Wang <wangyichao.sjtu.edu.cn> wrote:
> Dear Amber administrator,
>
> When the user submit two single-card jobs to GPU node with two cards on
> our cluster, I found that actually the two jobs are running on the same
> card.
>
> Could you have any idea to refine some parameters to let these two
> single-card jobs running on two different cards on the node?
>
> Looking forward to your help.
>
> Best regards,
>
> Stephen.
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Ask yourselves, all of you, what power would hell have if those imprisoned
here could not dream of heaven?
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Nov 30 2015 - 23:00:03 PST
