[slurm-users] How to apply for multiple GPU cards from different worker nodes?

[Christopher Samuel]

On 4/16/19 1:15 AM, Ran Du wrote:

>        And another question is : how to apply for multiple cards could 
> not be divided exactly by 8？ For example, to apply for 10 GPU cards, 8 
> cards on one node and 2 cards on another node?

There are new features coming in 19.05 for GPUs to better support them. 
You could then submit a 5 task job requesting --gpus-per-task=2 to 
achieve that (4 tasks on the first node and 2 on the second).

 From the release notes:

https://github.com/SchedMD/slurm/blob/master/RELEASE_NOTES

  -- Add GPU scheduling options for salloc, sbatch and srun:
    --cpus-per-gpu, -G/--gpus, --gpu-bind, --gpu-freq, --gpus-per-node,
--gpus-per-socket, --gpus-per-task and --mem-per-gpu.

All the best,
Chris
-- 
   Chris Samuel  :  http://www.csamuel.org/  :  Berkeley, CA, USA