<GREETING>

Below are <CASE> that ran on <CLUSTER> (<PARTITIONS>) in the past <DAYS> days:

<TABLE>

The CPU utilization (CPU-Util) of each of the <NUM-JOBS> jobs is approximately equal to
100% divided by the number of allocated CPU-cores (100%/CPU-cores). This
suggests that you may be running a code that can only use 1 CPU-core. If this is
true then allocating more than 1 CPU-core is wasteful (these jobs may have
wasted <CPU-HOURS> CPU-hours). A good target value for CPU utilization is 90%
and above.

Please consult the documentation of the software to see if it is parallelized.
For a general overview of parallel computing:

    https://your-institution.edu/KB/parallel-computing

If the code cannot run in parallel then please use the following Slurm
directives:

    #SBATCH --nodes=1
    #SBATCH --ntasks=1
    #SBATCH --cpus-per-task=1

You will experience shorter queue times by allocating only 1 CPU-core per job.
In some cases this will also allow you run more jobs simultaneously.

You can check the CPU utilization of completed and actively running jobs by using
the "jobstats" command. For example:

<JOBSTATS>

Replying to this automated email will open a support ticket with Research
Computing.
