[slurm-users] Sharing a node with non-gres and gres jobs
steinbac at mpi-cbg.de
Tue Mar 19 21:45:05 UTC 2019
As you can perhaps see from our slurm.conf, we have task affinity or similar switches off. Along the same route, i also removed the core binding of the GPUs. That is why, I am quite surprised, that slurm doesn’t allow new jobs in. I am aware of the PCIe bandwidth implications of a GPU running on the wrong socket or of the interfere des of pinned jobs versus unpinned ones. But at this point, these appear to be higher order optimizations.
More information about the slurm-users