Open loveshack opened 2 years ago
I assume you mean CUDA MPS?
https://docs.nvidia.com/deploy/mps/index.html
Would you be able to elaborate on it's use? It is not part of CUDA I have experience with.
You wrote:
I assume you mean CUDA MPS?
Yes, sorry.
https://docs.nvidia.com/deploy/mps/index.html
Would you be able to elaborate on it's use? It is not part of CUDA I have experience with.
No more than the doc, I'm afraid. (The current interest is potentially packing multiple gromacs processes in a GPU, rather than MPI.)
My example for gromacs is at https://fx.srht.site/posts/2022-05-26-gromacs-mps.html
I've been discussing using MPS, which isn't mentioned. It's probably worth at least saying it's available.