Regroup 3D - no resources reserved and job fails

Troubleshooting
1

Dear all,

we tried to use the Regroup 3D job to group similar 3D classes.
We are using CS 4.4.1 on a cluster with SLURM.
Below is a part of the SLURM script from a Regroup 3D job:

# What follows is a simple SLURM script:

#SBATCH --job-name P33_J222
#SBATCH -n 0
#SBATCH --gres=gpu:0
#SBATCH -p gpu
#SBATCH --mem=0G
#SBATCH --time=2-00:00

You can notice that CS reserves zero CPU cores (-n 0), GPUs (--gres=gpu:0) and RAM (--mem=0G). SLURM fails immediately since the task requested zero resources.

Is this a bug in the CS submission system?

Best!
Matic

2

@eMKiso You encountered a bug :beetle: in the Regroup 3D job for which we plan to release a fix.

1 Like
3

Hi @eMKiso ,

As of CryoSPARC v4.5, released today, this bug with Regroup 3D has been fixed; new jobs built in v4.5 should not suffer this issue.

Best,
Michael

1 Like