I have a cluster with 8 cores per node. I want to run 7 SLURM jobs per node.
Each Node is configured with 8 CPUs in slurm.conf. From the SLURM FAQ, I tried
adding this line to slurm.conf (and restarted the slurm daemons):
SelectType=select/cons_res
However, when I submit jobs with srun, I get error messages:
srun: error: slurm_receive_msg: Zero Bytes were transmitted or received
I also tried added 'Shared=YES' to the Partition in the slurm.conf file and had
problems.
How can I configure slurm so that I can run multiple jobs per node?
