Re: [AMBER] gpu run

From: Kenneth Huang <kennethneltharion.gmail.com>
Date: Sat, 5 Sep 2015 21:12:05 -0400

Hi,

I think in general what you want to do is first check the corresponding
numbers of the GPUs with something like nvidia-smi and then set

export CUDA_VISIBLE_DEVICES=0
export CUDA_VISIBLE_DEVICES=1
export CUDA_VISIBLE_DEVICES=2

Where 0, 1 and 2 all represent the different GPU cards, and then run each
job in a different session that has a corresponding GPU visible. That said,
that's what I usually do if I'm running on a workstation, so I'm not very
sure if you how it'd carry over to your particular machine or if you' d
need root access for it (I don't think you should), but that's something
that's more machine specific.

Best,

Kenneth


On Sat, Sep 5, 2015 at 7:32 PM, Lara rajam <lara.4884.gmail.com> wrote:

> Dear Amber
>
> I am running my jobs in multi gnu machine ( it has 3 GPU cards) .
> If I want to use a single GPU for each job so that I can use 3 gpu for 3
> different job
> as a user how can i give that and how can i check that. Since I am not
> aware of root password
>
> thank you
> _______________________________________________
> 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 Sat Sep 05 2015 - 18:30:03 PDT
Custom Search