Creating an HPC Cluster
For project spaces that support 2024 R1 and earlier Ansys releases
A cluster consists of several compute servers that are networked together. Each server is called a node. The nodes work in parallel with each other, boosting processing speed to deliver high performance computing.
HPC clusters can be created in project spaces that support 2024 R1 and earlier releases of Ansys applications. If you want to create a cluster with 2024 R2 or newer releases of Ansys applications, see Creating an Autoscaling Cluster instead.
You can add an HPC cluster to a project space if you are a tenant administrator or have Admin permission in the project space.
Configuring and connecting to an HPC cluster differs for each Ansys application. For detailed information, refer to the following topics in the Recommended Usage Guide:
For a quick reference on the hardware to choose for each application, see Recommended Hardware for HPC Workflows.
When a specific cluster package is deployed for the first time in a location, a dummy virtual machine is set up where the application packages for the cluster are installed. Afterward, an image is created from the dummy virtual machine, which is then used to create the cluster. This process can take anywhere between 20 minutes and 2 hours depending on the size and complexity of the cluster package. Clusters that are subsequently created in the availability zone generally take only 5 to 15 minutes to set up because the image has already been created. Note that these are rough estimates. Creation time varies depending on a number of factors, some of which may be related to AWS.
When the cluster is in the Running state, users can submit jobs to it. See Submitting a Job to an HPC Cluster in the User's Guide.