NVIDIA/nvidia-dockerArchived
Nvidia Docker
NVIDIA Docker is a container runtime wrapper that enables the use of host-level graphics processing units within isolated container environments. It functions as a containerized GPU orchestrator, mapping physical hardware resources into virtualized environments to support high-performance computing and machine learning workloads.
The project provides a toolkit that facilitates integration between containerized applications and host-level graphics hardware. By utilizing a pre-start hook to intercept container creation, the runtime injects necessary device drivers and libraries into the isolated environment, ensuring that graphics calls are redirected to host-provided drivers. This approach maintains compatibility with standard container engines and orchestration platforms by adhering to the Open Container Initiative runtime specification.
This infrastructure supports the deployment of hardware-accelerated computing tasks, including machine learning model training and scientific simulations. It manages the visibility of hardware resources by selectively exposing specific device nodes to the container process, allowing for consistent execution of complex software across different systems.
Features
- Containerized GPU Acceleration - Provides direct access to host graphics processing units for hardware acceleration within isolated containers.
- GPU Container Integration - Connects host graphics processing units to isolated container environments for hardware-accelerated computing.
- GPU Runtime Wrappers - Wraps container runtimes to automatically expose host graphics processing units for hardware-accelerated tasks.
- GPU Container Toolkits - Provides libraries and utilities that enable seamless integration between containerized applications and host-level graphics hardware.
- Cloud Native GPU Orchestration - Scales artificial intelligence applications by managing GPU resources through standard container orchestration.
- Infrastructure and Deployment - Docker container runtime for GPU-accelerated applications.
- 基础设施管理 - 针对 GPU 加速工作负载的容器运行时集成。
- Scientific Container Environments - Standardizes research environments by packaging complex simulation software and drivers into portable containers.
- Graphics Call Interceptors - Intercepts standard graphics calls and redirects them to host-provided drivers.
- Machine Learning Training - Supports the deployment of deep learning frameworks within containerized environments for model training.
- Driver Mapping Utilities - Maps host-side graphics device nodes and shared libraries directly into the container namespace.
- Hardware Device Exposure - Manages hardware visibility by selectively exposing host device nodes to container processes.
