This Repository contains a minimal example to build a docker container which can be used with nvidia-docker. A more detailed look into how it works can be found on my blog.
A small test program (which is built using program.nix) is wrapped into a docker container (using the generic builder
in buildCudaImage.nix).
To build the containers you'll need to get the nix build system.
Afterwards call nix-build in the project dir.
As this project demonstrates how to build nvidia-docker container, you need to have it installed. Your local CUDA
Version should be at least 10.0.130. If you require different version you can change the toolkits list in default.nix.
When all requirements are satisfied you can run all containers using ./result/run-script.sh.
Example output:
f9f30553cf96: Loading layer [==================================================>] 39.91MB/39.91MB
The image awesome-cuda-10.0.130:latest already exists, renaming the old one with ID sha256:7009448038a9fc1fdf80a3b467f3c4ddb502a32440f199cdf41f3a0dc5619e74 to empty string
Loaded image: awesome-cuda-10.0.130:latest
awesome-cuda-10.0.130
Runtime version 10000; Cuda error: 0 (no error)
Driver version 10010; Cuda error: 0 (no error)
Device count 1; Cuda error: 0 (no error)
1228318216db: Loading layer [==================================================>] 39.52MB/39.52MB
The image awesome-cuda-9.1.85.3:latest already exists, renaming the old one with ID sha256:c3006f9cc77298d438321d373ada218bf65dbd608509850bcebcaba2dbbcb737 to empty string
Loaded image: awesome-cuda-9.1.85.3:latest
awesome-cuda-9.1.85.3
Runtime version 9010; Cuda error: 0 (no error)
Driver version 10010; Cuda error: 0 (no error)
Device count 1; Cuda error: 0 (no error)
4c0fbc75e669: Loading layer [==================================================>] 39.8MB/39.8MB
The image awesome-cuda-9.2.148.1:latest already exists, renaming the old one with ID sha256:cafbd79f87cb7ac0e4bdad8d300f1c2499b1999f40ca1ea97e2f1fb8a1b72ffd to empty string
Loaded image: awesome-cuda-9.2.148.1:latest
awesome-cuda-9.2.148.1
Runtime version 9020; Cuda error: 0 (no error)
Driver version 10010; Cuda error: 0 (no error)
Device count 1; Cuda error: 0 (no error)