HelloWorld HPC Cluster Tutorial

This simple tutorial will walk you through running your first HPC job using SSH and SLURM job scheduler.

---

1. Create helloworld.py on your local computer using your favorite text editor. TextEdit and Notepad are two good options.

Create a Python script named helloworld.py with the following contents:

# helloworld.py
import time

for i in range(10):
    print(f"Hello from HPC! Iteration {i+1}")
    time.sleep(120)  # Sleep for 120 seconds so we can see the job in the queue

---

2. Upload helloworld.py to the cluster

Use scp:

scp helloworld.py your_utln@login.pax.tufts.edu:~/

Or upload using Open OnDemand, if available.

---

3. SSH into the cluster

ssh your_utln@login.pax.tufts.edu

---

4. Create a SLURM batch file using nano or vi

Run:

nano helloworld_job.sh

Paste this content:

#!/bin/bash
#SBATCH --job-name=hello_job
#SBATCH --output=hello_output.log
#SBATCH --error=hello_error.log
#SBATCH --time=00:02:00
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1

module load python  # Load Python if required by your system
python helloworld.py

Save and exit with Ctrl+O, Enter, Ctrl+X.

---

5. Submit your batch job

sbatch helloworld_job.sh

---

6. Check job status

squeue --me

---

7. Check output and error logs

After the job completes, view logs:

cat hello_output.log
cat hello_error.log

---

8. Check job efficiency with seff

Find your job ID and run:

seff JOB_ID

Example:

seff 123456

---

✅ You’ve now run your first HPC job using SLURM and SSH!