UCR Home ITS HPCC Campus status Get help
Research ComputingUniversity of California, Riverside
Home / Knowledge Base / Launching an Ursa Major HPC cluster
Cloud

Launching an Ursa Major HPC cluster

This guide is for a lab that wants to build its own Slurm cluster in its Ursa Major project with Google's Cluster Toolkit.

Before you build

  • Costs: a lab's own cluster is Tier 2. Its compute nodes, GPUs, disks and file systems are recharged to a lab funding source under an MOU. See KB005: Ursa Major service tiers and KB007: Ursa Major recharge workflow.
  • Exotic hardware: if you need hardware the HPCC does not have, such as TPUs or Arm processors, ask about Research Computing's shared cluster in Google Cloud instead. That is Tier 1, within limits set per project, and starts with a consultation.
  • The HPCC first: for most batch and GPU work, the campus HPCC cluster is a better fit and usually costs the lab less.
  • Set a budget on the project before you start. See Creating GCP budgets.

Talk to Research Computing before you build. We can help you compare options and estimate costs. For background, see Ursa Major HPC clusters.

Steps

Google's Cluster Toolkit quickstart: deploy a Slurm cluster walks through each step. Read its Costs section first.

  1. Select your lab's Ursa Major project in the Google Cloud console. If you do not have a project yet, see KB021: Requesting an Ursa Major (Google Cloud) project.
  2. Launch Cloud Shell.
  3. Make sure the default Compute Engine service account is enabled.
  4. Install Cluster Toolkit.
  5. Create the cluster deployment folder.
  6. Deploy the cluster with Terraform.
  7. Run a job on the cluster. For basic Slurm commands, see Connecting to the HPC cluster.
  8. Clean up and destroy the cluster when you no longer need it. Resources left running continue to be charged.

Video

Owner: Research Computing Reviewed: 4 Oct 2026
Related guides
GuideAlphaFold on a Slurm cluster with Google Cloud FilestoreGuideDistributed PyTorch training with Kubeflow TrainerKB002When your Ursa Major project is readyKB005Ursa Major service tiersKB007Ursa Major recharge projects: how setup worksKB021Requesting an Ursa Major (Google Cloud) projectGuideLLM Inference SettingsGuideRunning open-source language models with OllamaGuideUrsa Major guidelinesGuideUrsa Major HPC clustersGuideUrsa Major project budgets and spending alertsGuideUrsa Major research servicesGuideUrsa Major research workstationsGuideConnecting to an Ursa Major research workstationGuideLaunching an Ursa Major research workstation