Go to file
Felix Abecassis e7b792253f Allow images to be run using their hash value
Use "docker inspect" instead of "docker images".
2016-01-22 15:14:45 -08:00
centos-6/cuda Volumes now include drivers versioning to address drivers upgrade 2016-01-09 10:44:55 +01:00
centos-7/cuda Volumes now include drivers versioning to address drivers upgrade 2016-01-09 10:44:55 +01:00
mk Fix build 2016-01-09 10:44:55 +01:00
samples Add build variable to suffix pushed images with the OS version 2016-01-09 10:44:55 +01:00
tools Allow images to be run using their hash value 2016-01-22 15:14:45 -08:00
ubuntu-14.04 Volumes now include drivers versioning to address drivers upgrade 2016-01-09 10:44:55 +01:00
.gitattributes Enable folder "tools/" for language stats on GitHub 2016-01-10 14:20:02 -08:00
centos Add CUDA 7.0 and CUDA 7.5 Dockerfiles for CentOS 7.1 2015-11-17 16:56:18 -08:00
CLA Update CLA 2015-11-13 10:05:15 -08:00
LICENSE Bump copyright year 2016-01-04 19:53:16 -08:00
Makefile Do not put dependencies on the "push" rule, only push the current images 2016-01-09 10:44:55 +01:00
README.md Automatically pull images if they are not present locally 2016-01-11 17:25:56 -08:00
ubuntu Initial commit 2015-11-04 13:00:25 -08:00

NVIDIA Docker

This repository includes utilities to build and run NVIDIA Docker images.

nvidia-gpu-docker

Example of how CUDA integrates with Docker

Documentation

The full documentation is available on the repository wiki.
A good place to start is to understand why NVIDIA Docker is needed in the first place.

Quick start

Assuming the NVIDIA drivers and Docker are properly installed (see installation):

git clone https://github.com/NVIDIA/nvidia-docker

# Initial setup
sudo make install
sudo nvidia-docker volume setup

# Run nvidia-smi
nvidia-docker run nvidia/cuda nvidia-smi

Issues and Contributing

A signed copy of the Contributor License Agreement needs to be provided to digits@nvidia.com before any change can be accepted.