Skip to content

thejohnhoffer/sbatch_demo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Harvard Slurm Cluster

First you'll need to get access to the cluster.

This prints the singularity help from coxgpu05. As of 2017-09-13, coxgpu05 is the only node on the cluster that supports singluarity. Singularity can import docker images. Even with an RC Account, you will need to be added to the cox_lab group in order to use this machine.

Run bash singularity/test.sh

  • Outputs go to singularity/0.out
  • Errors go to singularity/0.err

This prints the documentation for pytorch from 4 different CPUs on 1-4 different cluster nodes. We create and use a conda virtual environment for python. We use sbatch --array to set up 4 parallel jobs differing only by $SLURM_ARRAY_TASK_ID. Full documentation for the --array option is here.

Run bash pytorch/test.sh

  • Outputs go to pytorch/*.out
  • Errors go to pytorch/*.err

Manage the jobs

Run this to check the status of the jobs:

squeue

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published