Autoscaling Clusters

A cluster consists of multiple machines that are networked together. Each machine is called a node. A cluster consists of a head node that handles the job scheduling, and worker nodes that do the work. The worker nodes work in parallel with each other, boosting processing speed to deliver high performance computing.

An autoscaling cluster is a cluster that dynamically scales the number of nodes up or down to meet the demand of the workload. When jobs are submitted to the cluster, nodes are spun up, and as jobs finish, the nodes are shut down.

A cluster can have a set number of static nodes that are allocated right away and remain available throughout the life of the cluster. Otherwise, nodes are allocated dynamically (as needed) to execute jobs submitted to the cluster. To control costs, you can put a limit on the number of dynamic nodes that can be spun up.

Only one autoscaling cluster can be present in a project space at one time. However, the cluster can have multiple job submission queues, and each queue can have a different application associated with it.

Autoscaling clusters can only be created in project spaces that support Ansys 2024 R2 and newer releases. If you want to create a cluster using Ansys 2024 R1 or earlier releases, see HPC Clusters instead.

Topics:

For step-by-step examples of autoscaling cluster configuration and usage, see Recommended Configurations by Application in the Recommended Usage Guide.

For general information on using autoscaling clusters, see Autoscaling Clusters in the User's Guide.

Important: If your company set up Ansys Gateway powered by AWS prior to March 4, 2025, the tenant administrator must update the permissions given to the shared access role before autoscaling clusters can be created. See Updating Account Permissions to Allow the Creation of Autoscaling Clusters in the Administration Guide.