Public | Automated Build

Last pushed: 2 years ago
Short Description
Short description is empty for this repo.
Full Description

This is Ubuntu 14.04 x86_64 LTS with Cuda 6.5 installed.

1) Make sure the driver version of host is 340.29. This is the one coming with CUDA 6.5.14.

The host must have the same version of Cuda driver installed. The image uses NVIDIA-Linux-x86_64-340.29 now. You can download this driver from

How to get my current Cuda driver version? Use:


2) Launch the container with GPU enabled.

(Special thanks to
Traun Leyden for pointing out the way to run CUDA images without starting docker in LXC mode.)

DOCKER_NVIDIA_DEVICES="--device /dev/nvidia0:/dev/nvidia0 --device /dev/nvidiactl:/dev/nvidiactl --device /dev/nvidia-uvm:/dev/nvidia-uvm"

docker run -ti $DOCKER_NVIDIA_DEVICES hjk41/ubuntu_14.04_with_cuda /bin/bash

If you have multiple GPU that you want to use in docker, you will need to modify the DOCKER_NVIDIA_DEVICES line and add more devices such as "--device /dev/nvidia1:/dev/nvidia1".

Any questions or suggestions, contact me at chuntao.hong (at)

Docker Pull Command
Source Repository

Comments (0)