RCC | Computing | VRL

Submit/Running Jobs

General Steps

The following is the general steps to run a job in the VRL cluster

  1. SSH to vrl.annauniv.edu
  2. Move your local files to your directory in the server
  3. Prepare a job submission script
  4. Submit the job via job submission script
  5. Check the status

Running on GPU nodes

#!/bin/bash
#SBATCH --job-name=newjob
#SBATCH --partition=gpu
#SBATCH --mail-user=hpcuser@annauniv.edu
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=2
#SBATCH --gres=gpu:1
#SBATCH --mail-type=ALL
#SBATCH --workdir=/home/hpcuser/deeplearning
#SBATCH --output=newjob_%j.out

cd $SLURM_SUBMIT_DIR
echo $SLURM_JOB_NODELIST > hostfile_$SLURM_JOBID

Running on CPU nodes

#!/bin/bash
#SBATCH -N 1
#SBATCH --ntasks-per-node=32
#SBATCH -J <testrun>
#SBATCH -p route
#SBATCH --time=24:00:00  
#SBATCH -o slurm.%N.%j.out # STDOUT
#SBATCH -e slurm.%N.%j.err # STDERR
#SBATCH --export=all
#SBATCH --mail-user=<username>@iiap.res.in
#SBATCH --mail-type=ALL 

cd $SLURM_SUBMIT_DIR
echo $SLURM_JOB_NODELIST > hostfile_$SLURM_JOBID

Running Python

Jupyter Lab, an integrated environment for python development is available at the URL

https://jlab.annauniv.edu

  • requires user registration to run codes on JLAB
  • alternatively, python can be run on the terminal after login to the server using ssh

Running Jupyter notebook

ssh -L 9999:localhost:9999 hpcuser@vrl.annauniv.edu