Currently, all jobs are running with the same default limits defined by our slurm configuration and some hard-coded sbatch parameters. At the least, we need to allow increasing the job runtime limit (currently 10 minutes) to support bigger jobs.
It's an open question of how/if users should be responsible for setting a job runtime limit, and how we set reasonable maximums to prevent overuse.
Currently, all jobs are running with the same default limits defined by our slurm configuration and some hard-coded sbatch parameters. At the least, we need to allow increasing the job runtime limit (currently 10 minutes) to support bigger jobs.
It's an open question of how/if users should be responsible for setting a job runtime limit, and how we set reasonable maximums to prevent overuse.