NVIDIA AI Enterprise and NVIDIA vGPU for Compute

NVIDIA vGPU for Compute enables multiple virtual machines to share a single physical GPU while providing compute capabilities for AI model training, fine-tuning, and inference workloads.

About NVIDIA vGPU for Compute

NVIDIA AI Enterprise is a cloud-native suite of software tools, libraries, and frameworks for production AI deployments. It delivers optimized performance, security, and stability with enterprise-grade support. NVIDIA AI Enterprise consists of two primary layers: the application layer and the infrastructure layer.

NVIDIA vGPU for Compute is licensed exclusively through NVIDIA AI Enterprise. NVIDIA vGPU for Compute distributes GPU resources efficiently across multiple VMs, optimizes utilization, and lowers overall hardware costs. It offers advanced monitoring and management capabilities including Suspend/Resume, Live Migration, and Warm Updates for Cloud Service Providers (CSPs) and organizations that require scalable, cost-effective GPU acceleration.

Documentation Structure

This documentation is organized into the following sections:

NVIDIA vGPU for Compute Documentation

• NVIDIA vGPU for Compute Overview
  • Glossary
  • NVIDIA vGPU Architecture Overview
  • NVIDIA vGPU for Compute Configurations
  • NVIDIA vGPU for Compute Key Features
  • Product Limitations and Known Issues
• NVIDIA vGPU for Compute Features
  • MIG Backed vGPU
  • Device Groups
  • GPUDirect RDMA and GPUDirect Storage
  • Heterogeneous vGPU
  • Live Migration
  • Multi-vGPU and P2P
  • NVIDIA NVSwitch
  • NVLink Multicast
  • Scheduling Policies
  • Suspend-Resume
  • Unified Virtual Memory (UVM)
• NVIDIA vGPU for Compute Installation
  • Prerequisites
  • Installing NVIDIA Virtual GPU Manager
  • Installing NVIDIA Fabric Manager on HGX Servers
  • Installing NVIDIA vGPU Guest Driver
• NVIDIA vGPU for Compute Licensing
  • Verifying the License Status of a Licensed NVIDIA vGPU for Compute Guest VM
  • Installing the NVIDIA GPU Operator Using a Bash Shell Script
  • Installing NVIDIA AI Enterprise Applications Software
  • Installing the NVIDIA AI Enterprise Software Components Using Podman
  • Installing NVIDIA AI Enterprise Software Components Using Kubernetes and NVIDIA Cloud Native Stack
• NVIDIA vGPU for Compute Configuration
  • Configuring MIG-Backed vGPU
• NVIDIA vGPU Types Reference
  • Frequently Asked Questions

Quick Start

New to NVIDIA vGPU for Compute? Follow this recommended path:

1. Understand the Basics - Start with Overview to learn key concepts, architecture, and vGPU configuration modes. For detailed feature information, refer to Features.

2. Install Components - Follow Installation to set up NGC CLI, Virtual GPU Manager, and Guest Drivers.

3. Configure Licensing - Set up Licensing to connect VMs to NVIDIA License System.

4. Configure vGPU - Use Configuration to set up MIG-backed vGPU and select vGPU profiles.

5. Reference vGPU Types - Consult Reference for complete vGPU type tables for your GPU.

Common Tasks

Installation and Setup

• Prerequisites - Verify hardware and BIOS settings.

• Installing NGC CLI - Download vGPU software from NVIDIA NGC.

• Installing Virtual GPU Manager - Deploy on hypervisor.

• Installing vGPU Guest Driver - Deploy in VMs.

Licensing

• Licensing vGPU VMs - Configure NVIDIA License System integration.

• Verifying License Configuration - Confirm licensing is working.

Configuration

• Configuring MIG-Backed vGPU - Enable MIG mode and create GPU instances.

• Virtual GPU Types - Select appropriate vGPU profiles for your workload.

Additional Resources

• NVIDIA AI Enterprise Product Support Matrix - Supported platforms and versions

• NVIDIA vGPU Software Documentation - Complete vGPU software documentation

• NVIDIA Fabric Manager User Guide - Multi-GPU configuration on HGX platforms

• NVIDIA NGC Catalog - Download vGPU software and drivers