AMD Resource Manager

AMD Resource Manager overview features

AMD Resource Manager#

This page summarizes what AMD Resource Manager is, how to begin, core capabilities, and essential terminology used across the documentation.

Solution overview#

The AMD Resource Manager provides administrators with tools to oversee and control computational resources and user access. Its key capabilities include cluster management, monitoring, and maintaining teams’ access to computational resources.

The AMD Resource Manager can be deployed standalone or combined with AMD AI Workbench. When deployed as standalone, the Resource Manager focuses on infrastructure and access control. When combined with AI Workbench, it adds enterprise-grade resource management to AI development workflows. See the Installation guide for deployment options and installation details.

Where to start#

Getting started

Install AMD Resource Manager, connect clusters, and complete your first setup steps.

Concepts

Core UI areas and how clusters, projects, workloads, and shared resources fit together.

Guides

Step-by-step tutorials for projects, quotas, workload monitoring, and resource sharing.

Key Features#

AMD Resource Manager is built around the basic usage pattern of maintaining compute resources, setting up teams and projects, and allowing individual users to utilize the resources for their compute needs.

  • Cluster: The physical part of the platform installation, which can be managed in the AMD Resource Manager user interface.

  • Organization: An organization is built from teams. Each team can have multiple users and multiple projects.

  • Projects: A project contains users and a quota for their workloads. Multiple users can belong to multiple projects.

  • Quota: A quota is a usage limit reserved for a project. Quotas are useful for ensuring everyone gets their fair share of compute resources.

  • Secrets: Secure information such as API keys or credentials that can be created at the organizational level and assigned to projects. Secrets ensure workloads can access what they need without exposing sensitive data.

  • User: Users are individuals who require compute access for work purposes.

  • Storage: Storage configurations provide the project with the required credentials and connection information for workloads to access storage options like S3. Like secrets storage configurations can be created at the organizational level and assigned to projects.

Terminology reference#

See the Glossary for terms used in project settings and hands-on guides.

Term

Explanation

GPU

A graphics processing unit. An essential part of compute clusters.

Node

A single data center computer that can contain multiple GPUs.

Cluster

A set of interconnected computational nodes.

Project

A container for AI development. The project is allocated resources through a quota.

Quota

A minimum set of guaranteed resources, ensuring fair allocation between projects.

SSO

Single sign-on. A user login feature allowing easy traversal between multiple applications.
Contents