---
title: "What is the HPC Cluster?"
canonical: "https://kb.uconn.edu/space/SH/27970568262/What%20is%20the%20HPC%20Cluster%3F"
format: markdown
---
#### Brief Overview

The Storrs HPC is a shared resource between all researchers on all UConn campuses. This guide will familiarize you with the need-to-knows of HPC before you get into the action.

#### Why use High Performance Computing?

Personal computers such as laptops or desktops are great for working, smaller file storage, and non-intensive computational tasks. They are quite flexible and sometimes portable for everyday use.

There are also more powerful machines that may exist in your lab or office. They typically have more CPUs, more memory, larger storage, and perhaps a powerful graphics card. These computers allow for more intensive tasks that common computers would struggle with.

A high-performance computing cluster has a collection of powerful computers that are networked together. It uses software that allows for coordinated running of jobs across multiple computers to accomplish tasks much faster. Additionally, since these computers are not running your desktop, you are not tied down and are free to continue working while your job runs in the background. The principal advantage of HPC is the ability to utilize multiple streamlined computers to run computations without holding your local machine hostage. 

#### On HPC

- Access is through a command line interface (CLI)
- Users do not communicate directly with computational nodes. Instead, they connect to a **login node**
- Jobs are submitted through a resource manager (SLURM) that allocates compute nodes

> 📝 **Definitions**
> 📝 
> 📝 **CPUs:** Central Processing Units or processors. CPUs on the HPC have multiple cores
> 📝 
> 📝 **Cores:** Basic computational unit of the CPU. Each core usually handles a single process at a time unless there is “hyper-threading” which allows for multiple processes to simultaneously be processed on the same core.
> 📝 
> 📝 **Node: **Can refer to a complete computer system that has multiple CPUs, memory, power supply, networking devices and local disk space.
> 📝 
> 📝 There are two types of nodes on the HPC:
> 📝 
> 📝 **Login nodes**: Nodes used to connect to and perform basic tasks on the HPC
> 📝 
> 📝 **Compute nodes**: Working nodes for computations. Some are CPU-only, others have GPUs

Once you’re on the HPC, there are a few things you’ll need to be familiar with to get working. Some of the most important of these are the following:

> ℹ️ #### Skills
> ℹ️ 
> ℹ️ 1. Proficiency in Linux command line
> ℹ️ 2. Knowledge of available hardware
> ℹ️ 3. Understanding of the SLURM Scheduler
> ℹ️ 4. Familiarity with the [Knowledge Base](https://kb.uconn.edu/space/SH/26024084708/Storrs+HPC)

In the next few guides, we’ll discuss these topics and log in to the Storrs HPC!

> ℹ️ **Next:** [https://uconn.atlassian.net/wiki/spaces/SH/pages/27970109455](https://uconn.atlassian.net/wiki/spaces/SH/pages/27970109455)