Posted On: Nov 2, 2022

AWS ParallelCluster 3.3 is now generally available and introduces a new feature for compute resource optimization. With this new feature, you can map a compute resource to a list of Amazon EC2 instance types with an allocation strategy to optimize compute capacity for your HPC jobs. Other features include updates that support dynamically mounting shared storage, Slurm accounting, and Amazon EC2 on demand capacity reservations (ODCR).

Significant feature enhancements in AWS ParallelCluster 3.3 include:

Multiple Instance type allocation: You can map compute resources on a queue to a list of Amazon EC2 instance types for greater flexibility in aggregating compute capacity for your HPC jobs.

Dynamic file system mounts:  You can mount file systems as shared storage onto a running cluster with little or no disruption to cluster operations and running jobs.

Support for Slurm accounting: You can create clusters with Slurm accounting enabled to manage and monitor your cluster, and push Slurm metrics to your database.

Support for On-Demand Capacity Reservations: You can configure a cluster to use Amazon EC2 On Demand Capacity Reservations (ODCR) while provisioning compute capacity.

AWS ParallelCluster is available at no additional cost, and you pay only for the AWS resources needed to run your applications. Learn how to launch an HPC cluster using AWS ParallelCluster here.

For more detail you can find the complete release notes for the latest version of AWS ParallelCluster here.