Submitting Jobs at HPC Facility (Slurm)

Jobs on the HPC cluster must be submitted using Slurm scripts.
Under NO circumstances should you execute your job directly on the Head node or any compute node.
All the jobs are to be submitted via the scheduler, not directly.

Job Scripts

Model scripts for submitting jobs:

Serial job: Serial_script.sh
Parallel job: OpenMPI_script.sh, OpenMP_script.sh
GPU job: Cuda_script.sh
Gaussian: gaus_script.sh

Submitting Jobs

sbatch <your_script.sh>: Submit a job to the default partition
sbatch -p <partition_name> <your_script.sh>: Submit a job to a specific partition
srun --pty bash: Submit an interactive job (you will get a shell)
scancel <jobid>: Cancel a job

Checking Job Status

squeue -u <username>: List jobs for a particular user
squeue -p <partition_name>: List jobs in a specific partition

Other Job-Related Commands

scancel <jobid>: Stop a running job
scontrol show nodes: List status of all compute nodes
sinfo -s: List available partitions and node info

Instructions for Gaussian Runs

Gaussian users need access to the Gaussian package (request at hpc@iitmandi.ac.in).
Login with X forwarding:
ssh -Y username@10.8.1.17
Add to your ~/.bashrc:

export PATH=$PATH:/opt/soft/share/g09
export PATH=$PATH:/opt/soft/share/g09/gv
export GAUSS_EXEDIR=/opt/soft/share/g09
export g09root="/opt/soft/share/g09"
export GV_DIR="${g09root}/gv"

To access ghostview:
gv

Gaussian Submit Script Example

Number of threads in Gaussian input (%nprocshared) must match Slurm CPU request.
Example for 4 threads:

#SBATCH --cpus-per-task=4
#SBATCH --job-name=gaussian_job
#SBATCH --output=gaussian_%j.out
#SBATCH --partition=day


srun g09 input.com

Submit using:
sbatch gaus.sh

Module Load Instructions

Before running any job, load required environment modules. Examples:

module avail          # List all available modules

Mail your suggestions/comments to it_helpdesk@iitmandi.ac.in