What is AWS Parallel Computing Service?
AWS Parallel Computing Service (AWS PCS) is a managed service that lets you run and scale high performance computing (HPC) workloads on AWS using Slurm. AWS operates the Slurm controller, while the compute nodes run as EC2 instances in your own AWS account.
The service combines compute, storage, networking, and visualization tools into an elastic environment. Users work in a familiar Slurm environment without having to operate the underlying infrastructure themselves.
Core Features
- Managed Slurm controller: AWS operates and updates the scheduler, including managed updates
- Compute node groups: Groups of elastically scaling EC2 instances that run jobs; login nodes can run as a dedicated node group
- Queues: Job queues that dispatch work to specific compute node groups
- Shared storage: Integration of Amazon EFS as a home directory and Amazon FSx for Lustre as a shared high performance file system
- Built-in observability: Observability features included for cluster operations
- Provisioning as code: Setup via the console, AWS CLI, AWS SDK, or AWS CloudFormation
Typical Use Cases
Tightly coupled MPI applications: Simulations in computer aided engineering, weather modeling, or seismic processing that communicate across many nodes.
High throughput computing: Large numbers of independent jobs, such as Monte Carlo simulations or genomics pipelines.
Accelerated computing: Workloads on GPUs, FPGAs, or AWS custom silicon.
Interactive workflows: Interactive access through login nodes for analysis and visualization.
Benefits
- Familiar Slurm environment without operating your own scheduler
- Elastic scaling of compute nodes through EC2
- Compute nodes run in your own AWS account and VPC
- Provisioning as infrastructure as code through CloudFormation
- Available in several EU regions
Integration with innFactory
As an AWS Reseller, innFactory supports you with AWS PCS: sizing the cluster controller and compute node groups, connecting Amazon FSx for Lustre and Amazon EFS, network and IAM design, and automating cluster provisioning with CloudFormation. We take cost control and the choice of an EU region into account.
Typical Use Cases
Technical Specifications
Frequently Asked Questions
Which scheduler does AWS PCS use?
AWS PCS uses Slurm. The Slurm controller runs in a service-owned AWS account, while the compute nodes run as EC2 instances in your own account. Users work with the familiar Slurm commands.
What are the building blocks of a PCS cluster?
A cluster provides the Slurm controller. Below it you define compute node groups, which are groups of elastically scaling EC2 instances, and queues that send jobs to those node groups. Login nodes can be set up as a dedicated node group.
How is AWS PCS billed?
According to AWS, usage incurs two separate charges: an hourly cluster controller fee that depends on the controller size, and an hourly node management fee for every EC2 instance launched in your compute node groups. Optional Slurm accounting is billed separately. The EC2 instances, storage, and data transfer you use are charged additionally at standard rates.
Is AWS PCS available in the EU?
Yes. The documentation lists endpoints in Frankfurt (eu-central-1), Ireland (eu-west-1), London (eu-west-2), Milan (eu-south-1), Paris (eu-west-3), Spain (eu-south-2), and Stockholm (eu-north-1), among others. See the official documentation for the current list.
Which shared storage does AWS PCS support?
The AWS tutorials attach Amazon EFS as a shared home directory and Amazon FSx for Lustre as a shared high performance directory. Refer to the official documentation for further options.
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of AWS (official documentation). This page does not represent an offer by AWS.