We are deploying 2 compute nodes with Nvidia v100 GPUs and would like to use the CUDA MPS feature. I am not sure as to where to get the number to use for mps when defining the node in the slurm.conf?
Any advise would be greatly appreciated. Regards, SS
