[Hps] Recent Farm Job Memory Efficiency

[sciadm at jlab.org]

Jobs running on the farm are allocated memory for their exclusive use. Jobs that request significantly more memory than they use create a memory shortage and leave portions of the farm idle.

You are receiving this email because at least 284 of your recently completed jobs were memory-inefficient. These jobs used less than 0.09% of the requested memory. Reference https://scicomp.jlab.org/scicomp/slurmJob/memoryEff?user=hps for detailed job memory information. Please consider your memory request to improve throughput for yourself and others.

For overall workflow planning, the farm is provisioned for 2GB of memory per CPU.Jobs that exceed this will cause some portion of the farm to be idle even if they are memory-efficient.