HPCC

A shared high-performance computing cluster for research, creative work, and coursework.

Hardware

Node type	Count	CPU cores	Memory	GPUs
`gpu`	[N]	[N]	[N] GB	NVIDIA H100 / A100
`cpu`	[N]	[N]	[N] GB	—

Run sinfo on the cluster for the current partition and node list.

Accounts

Accounts are sponsored by a faculty PI. Contact [support] to request access; students join a PI’s project rather than requesting their own.

Logging in

Connect over SSH from the campus network or VPN:

ssh <username>@hpcc.cs.depaul.edu

Storage

Path	Purpose	Quota	Backed up
`/home/<user>`	Source code, small files, configs	[15 GB]	Yes
`/scratch/<user>`	Job input/output, large data sets	[large]	No — purged after [N] days

Running jobs

Do not run computations on the login node — it is shared by everyone. Submit work to the SLURM scheduler with sbatch. A minimal GPU job script:

#!/bin/bash
#SBATCH --job-name=train
#SBATCH --partition=gpu
#SBATCH --gres=gpu:a100:1
#SBATCH --cpus-per-task=8
#SBATCH --time=04:00:00

module load cuda
srun python train.py

Submit it with sbatch job.slurm and check status with squeue --me.

Software

Software is provided as environment modules — list with module avail and load with module load. Containers run under [Apptainer]. Docker is not available on the cluster.

Support

Email [support address] with your username, the job ID, and the error output. Office hours are held [day, time].

Citing the HPCC

If the HPCC contributed to published work, please acknowledge it. Suggested wording:

Computations were performed on the High-Performance Compute Cluster operated by the Jarvis College of Computing and Digital Media at DePaul University.