User Guide¶
This guide is for people who use DC Suite to run GPU workloads — data scientists, ML engineers, researchers, and their teams. You do not need to know how the platform is deployed; you just need an account.
What you can do¶
| Task | Where |
|---|---|
| Sign in and find your way around | Getting Started |
| Understand clusters, profiles, images, tenants | Core Concepts |
| Spin up a GPU cluster | Creating a Cluster |
| Resize, power-cycle, or delete a cluster | Managing Clusters |
| SSH in and expose web UIs | Access & Networking |
| Understand storage tiers and keep your data | Storage & Data |
| Submit and track SLURM jobs | Running Jobs |
| Apply software stacks to a cluster | Templates |
| Build and promote machine images | Images |
| Watch live metrics and logs | Observability |
| See what you are spending | Cost & Billing |
| Understand quotas and availability | Quotas & Limits |
| Understand isolation and access control | Security & Isolation |
| Automate against the REST API | API Reference |
Prerequisites¶
- An account in your organization's DC Suite installation (created by an administrator and signed in via your identity provider).
- A role that grants the actions you want to perform. If a button is missing or an action is refused, you likely lack the permission — ask an admin.
- For SSH access: an SSH key pair on your machine.