Here's another output: # squeue -o "%10F %10c %10C %10T %10M %10R" ARRAY_JOB_ MIN_CPUS CPUS STATE TIME NODELIST(R 59 1 1 PENDING 0:00 (Resources 60 1 1 PENDING 0:00 (Priority) 61 1 1 PENDING 0:00 (Priority) 62 1 1 PENDING 0:00 (Priority) 63 1 1 PENDING 0:00 (Priority) 64 1 1 PENDING 0:00 (Priority) 65 1 1 PENDING 0:00 (Priority) 66 1 1 PENDING 0:00 (Resources 57 1 8 RUNNING 1:21 node213 58 1 8 RUNNING 1:18 node236
On Tue, Jul 12, 2016, at 15:57, Yuri wrote: > Forgot to say that only the last two lines are the jobs that are > running. > > On Tue, Jul 12, 2016, at 15:55, Yuri wrote: > > > > - > > # squeue -o "%A %c %C" > > JOBID MIN_CPUS CPUS > > 48 1 1 > > 49 1 1 > > 50 1 1 > > 51 1 1 > > 52 1 1 > > 53 1 1 > > 46 1 8 > > 47 1 8 > > - > > > > Each job is the same bash script: > > - > > #!/bin/bash > > #SBATCH -n 1 > > #SBATCH -t 0-00:05 > > #SBATCH -p short > > > > sleep 200 > > - > > > > In slurm.conf I have CPUs=4 for each node (but each node actually has a > > Intel Core i7). My question is: why is slurm assigning only one job per > > node and each job is consuming 8 CPUs? This should not be happening > > because the script is single CPU and each node has only 4 CPUs in the > > specification, so the correct would be 4 jobs per node, right?
