Overview

To run artificial intelligence (AI), machine learning (ML), or high performance computing (HPC) workloads on Hypercompute Cluster, you can create VMs, managed instance groups (MIGs), Slurm or Google Kubernetes Engine (GKE) clusters. For more information about Hypercompute Cluster, see Hypercompute Cluster.

Note: Hypercompute Cluster is only supported for A3 Ultra machine types. If you want to use a different machine type, the guides in this section won't work. To identify the VM and cluster creation guide for other machine types, see Choose a deployment strategy.

To create your VMs or clusters, complete the following steps:

1. Review the available configurations for Hypercompute Cluster. This outlines the deployment type, scheduling type, and provisioning model for Hypercompute Cluster. It also provides recommendations for operating systems that are best suited for your infrastructure. See Review configurations.
2. Create a reservation for the A3 Ultra machine type. See Request capacity.
3. Create VMs, MIGs, or clusters.
   • To create VMs, you have the following options:
     • Create VMs
     • Create VMs using Bulk API
   • To create MIGs, see Create a managed instance group (MIG)
   • To create Slurm clusters, see Create a Slurm cluster
   • To create Google Kubernetes Engine clusters, you have the following options.
     • Create GKE clusters with default configurations, see Create a GKE Hypercompute Cluster with default configuration
     • Create GKE clusters with custom configurations, see Create custom GKE Hypercompute Clusters