submitit icon indicating copy to clipboard operation
submitit copied to clipboard

Published 20 hours ago • verified publisher icon facebookincubator

Error whilst

Open mathota opened this issue 3 years ago • 1 comments

Hi,

I was trying to run the example(add(a,b)) provided https://github.com/facebookincubator/submitit, my hpc cluster is throwing following error (script1) submitit.core.utils.FailedJobError: sbatch: error: Job rejected: Please do not specify cores/CPUs/tasks for GPU jobs.

So when I unset ntasks_per_node variable, my hpc cluster is throwing the following error (script 2) IndexError: tuple index out of range

Can you please advice to resolve this script1.txt script2.txt

mathota avatar Sep 22 '21 20:09 mathota

Hi Which version are you using? if you don't specify anything, cores/cpus/tasks should not be preset in the latest version (it's been changed recently). If you are using the latest version, can you please print the submission file corresponding to script1?

For the second option, it bugs because asking for no task is not really expected. Setting slurm_ntasks_per_node=None should work though, and it should also work after merging #1652

jrapin avatar Sep 27 '21 09:09 jrapin