drewfarris / docker-cuda-arm64v8

Dockerfiles for cuda on arm64v8 (Jetson TX2)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

CUDA Docker images for Jetson TX2 / L4T 27.1

This work builds upon the official nvidia cuda docker images and the tegra docker project to provide Dockerfiles to build cuda and cudnn images for the JetsonTX2 running L4T 27.1 (Ubuntu 16.04) as installed by JetPack 3

There is some hackery here, specifically exposing the local apt repos installed by Jetpack via an nginx server so that the docker build process can use the standard approach of installing packages via apt.

The 8.0/runtime directory requires a copy of deviceQuery in order to build successfully. See the tegra docker project readme for more information. Fundementally, this is used for validation that the containers can observe the proper devices when running so it's not a >strict< requirement that it be present.

CUDA 8.0

About

Dockerfiles for cuda on arm64v8 (Jetson TX2)


Languages

Language:Nginx 72.9%Language:Shell 27.1%