News

AWS Launches Managed Parallel Computing Service to Reduce HPC Overhead

Amazon's new Parallel Computing Service (PCS) became generally available recently, promising to alleviate much of the administrative burden associated with running high-performance computing (HPC) workloads.

AWS PCS is supported in the US East (Ohio), US East (N. Virginia), US West (Oregon), Europe (Frankfurt), Europe (Stockholm), Europe (Ireland), Asia Pacific (Sydney), Asia Pacific (Singapore) and Asia Pacific (Tokyo) regions.

The service is specifically designed for scientists and engineers who use the AWS cloud to run HPC instances for tasks such as drug discovery, genomic analysis, weather forecasting, and building engineering models. These workloads typically require outsized compute power to support complex simulations and large data volumes. They also benefit from running instances in parallel, allowing multiple computations with various dependencies to be performed simultaneously.

Creating and managing HPC workloads at scale, however, can be very resource-intensive. The new PCS solution aims to simplify the process.

"Managing HPC workloads, particularly the most complex and challenging extreme-scale workloads, is extraordinarily difficult," said Ian Colle, AWS's director of advanced compute and simulation, in a prepared statement.

"Our aim is that every scientist and engineer using AWS Parallel Computing Service, regardless of organization size, is the most productive person in their field because they have the same top-tier HPC capabilities as large enterprises to solve the world’s toughest challenges, any time they need to, and at any scale."

PCS lets users orchestrate their HPC clusters using the open source Slurm job scheduler. Users who are already using Slurm for their on-premises workflows can migrate them to the AWS cloud with zero architecting required.

Users can deploy new Slurm environments using the AWS Management Console, command-line interface and SDK. AWS also plans to eventually add support for its infrastructure-as-code tool, CloudFormation, to PCS.

More information on AWS PCS is available on this product page and this technical blog post.

About the Author

Gladys Rama (@GladysRama3) is the editorial director of Converge360.

Featured

Subscribe on YouTube