Frequently Asked Questions

Login is done via an SSH connection in your terminal of choice. Use the following command to login:

ssh username@wi-hpc

Where "username" is your personal username and when prompted for a password, use the same password as you use to unlock your computer.

If you cannot access with the command above, please submit ticket by sending an email to the IT Help desk at helpdesk@wistar.org to request access.

I just logged in for first time, how do I submit a job?

Jobs can be submitted via the commands srun and sbatch. Please see the Scheduling Jobs page for more information on how to submit jobs.

Why won't my job start?

The first thing to do is run squeue. This will show all running and pending jobs in the cluster. To find your specific job run squeue -u <USERNAME>. Please see the Monitoring Jobs page for more information on checking the status of jobs.

PLEASE NOTE that if you are submitting a job during, or near a maintenance period, your jobs may not run until the period is over. Please look at out for HPC Maintenance emails coming from the IT Department.

How do I get job information through email?

To receive emails whenever a job starts, ends, fails, or requeue's via email add the following to your submission script.

#SBATCH --mail-user=<USERNAME>@wistar.org
#SBATCH --mail-type=BEGIN
#SBATCH --mail-type=END
#SBATCH --mail-type=FAIL
#SBATCH --mail-type=REQUEUE
#SBATCH --mail-type=ALL

What is the maximum number of jobs I can submit?

The cluster allows a user to submit up to 10,000 jobs at one time. This limitation exists to avoid hanging the cluster scheduling system. If your requirements are >= 10,000 you should use job logic to manage these additional job submissions. One example of job logic is the use of job arrays, please see our Job Arrays section on our Scheduling Jobs page.

How long can my job run for?

If you do not specify a Wall Time in your job submission (i.e. --time=D-HH:MM) your job will default to 1 HOUR.

Times also depend on what partition/queue you are submitting your job to. The defq and gpu queue are unlimited, whereas the smq has a max of 4 HOURS.

Please see our section on Wall Times and Partitions for more detailed information.

How do external collaborators get an account?

To get an outside collaborators an account to access the cluster, please submit a ticket to the IT Helpdesk at helpdesk@wistar.org

Why can I not run X11 in my session?

X11 forwarding is not enabled by default when entering an interactive session. You must include the --x11 flag on your srun command.

srun --x11 --{{ other flags }}  /bin/bash

I keep getting an Out-of-Memory error (OOM), what do I do?

When running or submitting jobs ensure that you are specifying the amount of memory you wish to use. By default memory-per-cpu is not set and totally memory is 1GB. Please refer to the Submitting jobs section of our Job Scheduling page.