AWS is announcing the launch of AWS Parallel Computing Service (PCS), a new managed service that sets up and manages high performance computing (HPC) clusters.
HPC brings together multiple computers or resources to tackle more complex or resource-intensive workloads, like computational fluid dynamics, weather modeling, finite element analysis, electronic design automation, and reservoir simulations. With AWS PCS, customers can work in their familiar HPC environment without needing to worry about infrastructure.
AWS PCS uses the job scheduler Slurm for scheduling and orchestrating HPC clusters. System admins can create managed clusters through Slurm that use their compute and storage configurations, identity, and job allocation preferences, and then end users can run and manage their HPC jobs, use interactive software on virtual desktops, and access data. For customers interested in getting started, Amazon has put together a tutorial for creating a simple cluster in AWS PCS in its documentation.
This announcement follows AWS’ 2018 release of AWS ParallelCluster, which is also a management tool for deploying and managing HPC clusters in AWS. According to the company, customers have been asking for a managed service so that they wouldn’t need to manage updates, which requires tearing down and then redeploying clusters.
AWS PCS is accessible through the AWS Management Console, AWS SDK, and AWS CLI. In terms of billing, it launches all resources in a customer’s AWS account and will charge for those resources used, AWS explained.
It is currently available in the following AWS regions: US East (N. Virginia), AWS US East (Ohio), US West (Oregon), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), and Europe (Stockholm).
