We at the HPCf have decided to make some changes in our policies in order to improve the sharing of our computing resources among our users. Effective immediately, new jobs submitted to Nanobio will have a maximum allowed running time. That is to say, jobs will be allowed to run until a time limit is reached, at which point if they have not completed their run they will be killed by the queue manager.
When submitting a job you will now have two options:
- specify a maximum running time, or
- don’t specify a maximum running time and accept the default value.
If you do not specify a maximum running time, your job will run with the default maximum running time which will be set to 30 minutes. If the default running time is too short for your needs, you will need to specify a maximum running time yourself. To do so, you must add the following directive to your job script:
#$ -l h_rt=hh:mm:ss
where hh, mm, and ss represent the hours, minutes, and seconds you will request, respectively. For example, to request 12 hours you would add the following:
#$ -l h_rt=12:00:00
The maximum amount of time you may request is 1 week (168 hours). Any job that runs longer than 1 week will be automatically killed by the queue manager.
If your job requires a running time greater than 1 week, you must refer to the manual of the software you will be running to confirm what checkpointing capabilities it provides. Checkpointing allows jobs to be interrupted and then resubmitted to resume from where they left off.
We invite you to refer all your questions to help(at)hpcf.upr.edu. As always, we are glad to assist you in any way we can. Thank you for your continued support for and cooperation with the HPCf.
Note: This change in policy will affect all jobs newly submitted after this notice. Previously submitted jobs that are currently running will not be affected by this change.