I am working with a user who is trying to optimize his job. Each node has 64 cores and 529419740kb of memory. Here is the output of tracejob. I am guessing vmem isan indication of how much swaapping is taking place. He is asking for 320 cores or five nodes.
02/18/2021 23:28:27 S Exit_status=0 resources_used.cpupercent=6398 resources_used.cput=172:08:25 resources_used.mem=83939560kb resources_used.ncpus=320 resources_used.vmem=264742324kb