Pre-Installed GPU Container Runtime

Overview

GPU Container Runtime is a GPU-aware container runtime built on the Open Container Initiative (OCI) specification. Developed by NVIDIA, it provides a seamless bridge between Docker (or other container technologies) and NVIDIA GPUs, enabling containers to directly access and utilize host GPU resources without complex virtualization or manual configuration. It integrates the NVIDIA Container Toolkit to automatically manage GPU drivers, libraries, and device mounting, supporting multi-GPU configurations and compatibility with orchestration platforms like Kubernetes and Docker Swarm.

Select a pre-installed GPU Container Runtime Marketplace image and get a fully configured server that's ready to use immediately—no manual installation or complex setup required.

Supported Operating Systems

The GPU Container Runtime Marketplace image is available for the following Ubuntu operating systems (all 64-bit):

• Ubuntu Server 24 LTS 64-bit
• Ubuntu Server 22 LTS 64-bit
• Ubuntu Server 20 LTS 64-bit
• Ubuntu Desktop 22 LTS 64-bit

Supported Products / Plans

The GPU Container Runtime Marketplace image is currently available across 8 GPU VPS plans and 20 GPU-dedicated server plans:

GPU VPS Plan:

• Express GPU VPS - 2GB
• Basic GPU VPS - RTX 5060
• Professional GPU VPS - RTX Pro 2000
• Professional GPU VPS - RTX A4000
• Advanced GPU VPS - RTX Pro 4000
• and more

GPU-Dedicated Server Plans:

• Lite Dedicated GPU Server - P600
• Express Dedicated GPU Server - P1000
• Basic Dedicated GPU Server - T1000
• Basic Dedicated GPU Server - GTX 1650
• Basic Dedicated GPU Server - K80
• Basic Dedicated GPU Server - GTX 1660
• and more

View all compatible plans →

GPU Container Runtime Version

Database Mart offers the latest version of GPU Container Runtime, pre-installed and ready to use.

What's Installed

When you select the GPU Container Runtime Marketplace image, Database Mart automatically installs and configures:

Core Operating Environment

• Operating System: Ubuntu Server or Desktop (versions 20.04, 22.04, or 24.04 LTS), pre-optimized for GPU-accelerated container workloads
• GPU Drivers & Acceleration: NVIDIA GPU drivers pre-installed and verified with nvidia-smi
• Docker Engine: Docker CE or EE pre-installed and configured for immediate use

GPU Container Runtime Components

• NVIDIA Container Toolkit: Pre-installed and configured — the core tool that enables containers to access NVIDIA GPUs
• NVIDIA Container Runtime: The OCI-compliant runtime that acts as a bridge between Docker and NVIDIA drivers
• Docker Integration: The nvidia runtime is pre-configured as the default runtime in the Docker daemon configuration (/etc/docker/daemon.json)

Configuration & Management

• Multi-GPU Support: Pre-configured to support multiple GPUs in a single container or across multiple containers
• Device Access: GPU devices (/dev/nvidia*) are pre-configured for transparent mounting
• CUDA Library Management: CUDA driver and libraries are automatically shared with containers

Installation Process

During server provisioning, Database Mart automatically performs the following steps:

1. Deploys the selected Linux operating system (Ubuntu 20.04 LTS, 22.04 LTS, or 24.04 LTS).
2. Installs NVIDIA GPU drivers and verifies GPU availability using nvidia-smi.
3. Installs Docker Engine and configures the Docker daemon for GPU support.
4. Installs and configures the NVIDIA Container Toolkit, including:
   • Adding the NVIDIA Container Toolkit repository
   • Installing the nvidia-container-toolkit package
   • Configuring the nvidia runtime in the Docker daemon configuration (/etc/docker/daemon.json)
   • Restarting the Docker service to apply changes
5. Verifies GPU container access by running a test container with docker run --gpus all nvidia/cuda nvidia-smi.

Everything is fully automated—your GPU Container Runtime server is ready to use as soon as deployment is complete.

Getting Started After Deployment

After your server has been provisioned, follow these steps to begin using GPU Container Runtime.

Step 1: Access the Server

Once deployment is complete, connect to your server via SSH using the credentials provided by Database Mart:

ssh administrator@your-server-ip

Step 2: Verify the Installation

Check that all components are properly installed and the GPU is accessible:

# Verify GPU driver installation
nvidia-smi

# Check Docker version
docker --version

# Verify NVIDIA Container Toolkit is installed
nvidia-ctk --version

Step 3: Run a Test GPU Container

Run the official NVIDIA CUDA test container to confirm GPU access:

docker run --rm --gpus all nvidia/cuda:13.0.1-devel-ubuntu24.04 nvidia-smi

This command:

• Enables access to all GPUs (--gpus=all)
• Runs the NVIDIA CUDA development image
• Executes nvidia-smi to display GPU information

If GPU support is configured correctly, the command will display information about your NVIDIA GPU, including the driver version, CUDA version, and available GPU devices.

Step 4: Run a Container with Specific GPU Access

To target specific GPUs, use the device_ids field:

# Access only GPU-0 and GPU-3
docker run --rm --gpus '"device=0,3"' nvidia/cuda:13.0.1-devel-ubuntu24.04 nvidia-smi

Or use Docker Compose with GPU reservations:

services:
  test:
    image: nvidia/cuda:13.0.1-devel-ubuntu24.04
    command: nvidia-smi
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

Step 5: Run Your Own GPU-Accelerated Container

Once GPU access is verified, you can run any GPU-accelerated container:

# Run a TensorFlow container with GPU access
docker run --rm --gpus all tensorflow/tensorflow:latest-gpu python -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

# Run a PyTorch container with GPU access
docker run --rm --gpus all pytorch/pytorch:latest python -c "import torch; print(torch.cuda.is_available())"

Notes

Key Benefits:

• Seamless GPU Access: Containers can directly access host GPUs without complex virtualization
• Automatic Driver Management: GPU drivers and libraries are automatically managed by the NVIDIA Container Toolkit
• Multi-GPU Support: Run containers with access to all GPUs or specific devices
• Container Orchestration Compatibility: Works with Kubernetes, Docker Swarm, and Docker Compose

Frequently Asked Questions

Q: What is GPU Container Runtime and why do I need it?

GPU Container Runtime is a GPU-aware container runtime built on the Open Container Initiative (OCI) specification. Developed by NVIDIA, it provides a seamless bridge between Docker and NVIDIA GPUs, enabling containers to directly access and utilize host GPU resources without complex virtualization or manual configuration. It integrates the NVIDIA Container Toolkit to manage GPU drivers, libraries, and device mounting automatically.

Q: What operating systems are supported?

The GPU Container Runtime Marketplace image is available for Ubuntu Server 24.04 LTS, 22.04 LTS, and 20.04 LTS, as well as Ubuntu Desktop 22.04 LTS — all 64-bit.

Q: What is the NVIDIA Container Toolkit?

The NVIDIA Container Toolkit is the core component that enables GPU access in containers. It includes the NVIDIA Container Runtime and provides the necessary components to expose GPU devices and CUDA libraries to containerized applications. It replaces the older nvidia-docker2 package and is the current recommended solution.