Getting Started
Hello world
Welcome to the Storrs High-Performance Computing (HPC) facility! This page is your friendly starting point for launching your computational research here at UConn. With instructions on everything from connecting to the HPC to submitting jobs, our step-by-step guide is tailored to make your journey into high-performance computing as smooth as possible. We’ve also designed this page to be adaptable to meet the needs of both newcomers and experts.
Let's get started so you can take advantage of everything our HPC has to offer!
Step 1: Make an account
The first step to using the HPC is to request an account. If you don't already have an account, please fill out the cluster application form. It’s free!
Setting up your account can take a bit of depending on how busy the Storrs HPC Admins are. While you’re waiting for your request to be filled, consider reading through the General HPC Guidelines. We’ll email you when your account is all set up. Then you can proceed to Step 2.
Step 2: Are you connected to UConn-Secure Wifi or an on-campus ethernet port?
Yes
No
Step 3: What kind of computer do you have?
The instructions for logging into the HPC are slightly different depending on what kind of computer you have. Please select what kind you have from the below options:
Sometimes DUO for 2FA authentication can deny connections even though the netid password entered is correct. If this happens, it will cause a connection issue 3 times and will prevent the connection from being established to HPC. If this happens, we recommend resetting the connection and request the DUO push. Eventually DUO should allow the connection to HPC. We are aware of the issue and will decrease the 3 stage error check to 1 in a coming outage.
Windows/PC
Mac
Linux
Step 4: Have you used the “Unix shell” before?
No…. What is this “shell” you speak of…?
Yes, but I’d love a refresher.
Yes, I use the Unix shell and command line regularly.
Step 5: Have you used an HPC before?
No
Yes
Step 6: Does the command srun
mean anything to you?
No / I am familiar with the command but would like a refresher.
Yes, I have a long history of starting interactive jobs with various srun
options on HPCs.
Step 7: Are you familiar with how batch jobs work on HPCs?
I’m a beginner
Yes, but I’d like a brief refresher
Yes, I have significant experience using sbatch
to submit SLURM jobs
Step 8: Are you familiar with the modules
system for loading software?
No
I would like a refresher
Yes
Step 9: Summary of Resources Available on the Storrs HPC
Standard Cluster Nodes
There are two main classes of standard nodes (i.e., without graphics cards a.k.a GPUs) available on the HPC cluster.
Name | Nodes | Available cores per node** | Cores Total | RAM (GB) | Names |
---|---|---|---|---|---|
Epyc64 | 41 | 62 | 2,542 | 503 | cn410-447,453-455 |
Epyc128 | 148 | 126 | 18,648 | 503 | cn456-603 |
**2 cores are reserved per node for the OS and storage processes.
GPU Cluster Nodes
There are a variety of GPU nodes available on the Storrs HPC as well. Most of our GPU nodes are Epyc64 nodes with between 1 and 3 GPUs.
Partitions
Storrs HPC has several different “partitions,” groups of nodes with similar hardware and usage types. All users have access to general partitions, while additional access to priority nodes can be purchased. The HPC provides various computational setups (a.k.a. “architectures”) with different strengths too—such as high core count, ample RAM, or GPUs—allowing users to select the best hardware for their research.
Step 10: Closing Thoughts
Please read our usage policy and General HPC Guidelines.
If you’re new to HPCs, also consider checking out our glossary where we explain the components of an HPC and the basics of how SLURM works.
HPC applications
We have created helpful software guides to demonstrate how to effectively use popular scientific applications on the HPC cluster.
Troubleshooting
If you have trouble using the HPC, please view our troubleshooting guide where we explain how to document errors and get help from Storrs HPC Admins (i.e., “submit a ticket”). Another good stop is our Frequently Asked Questions (FAQ) page where we address many common errors.