This documentation references AWS ParallelCluster. For running HPC and AI workloads, we recommend AWS Parallel Computing Service (AWS PCS), our managed Slurm service that lets you focus on running your workloads rather than managing infrastructure.

Setting up AWS ParallelCluster

The following topics describe how to set up AWS ParallelCluster. You will learn how to install the necessary tools and how to use them, how to implement and manage multiple user access to clusters, and the best practices.

Prerequisites

Before you can start setting up and using AWS ParallelCluster, make sure that you've completed the following prerequisites.

Sign up for an AWS account

To get started with AWS, you need an AWS account. There are two ways to sign up for AWS. If you want complete control over account configuration and where your content is stored, or if you have regulated workloads such as HIPAA or FedRAMP, use Sign up for AWS (advanced). If you are a new user and want to use a login you already own to sign in to AWS and create an AWS environment with preconfigured defaults in select AWS Regions, use Sign up for AWS.

Create a key pair

To deploy clusters, AWS ParallelCluster launches Amazon EC2 instances to create the cluster head node and compute nodes. To perform cluster tasks, such as running and monitoring jobs, or managing users, you must be able to access the cluster head node. To verify you can access the head node instance using SSH, you must use an Amazon EC2 key pair. To learn how to create a key pair, see Create a key pair in the Amazon Elastic Compute Cloud User Guide for Linux Instances.