adding NVIDIA capabilities information

Signed-off-by: Adrian Plata <adrian.plata@docker.com>
This commit is contained in:
Adrian Plata 2019-07-02 15:48:30 -07:00
parent a322063892
commit d8aa3f275b
1 changed files with 19 additions and 0 deletions

View File

@ -332,3 +332,22 @@ $ docker run -it --rm --gpus device=0,2 nvidia-smi
Exposes the first and third GPUs.
> **Note**: NVIDIA GPUs can only be accessed by systems running a single engine.
#### Set NVIDIA capabilities
You can set capabilities manually. For example, on Ubuntu you can run the
following:
```
docker run --gpus 'all,capabilities=utility' --rm ubuntu nvidia-smi
```
This enables all available GPU driver capabilites with the driver features for
`nvidia-smi`.
More information on valid variables can be found at the [nvidia-container-runtime](https://github.com/NVIDIA/nvidia-container-runtime)
GitHub page. These variables can be set in a Dockerfile.
You can also utitize CUDA images which sets these variables automatically. See
the [CUDA images])https://github.com/NVIDIA/nvidia-docker/wiki/CUDA) GitHub page
for more information.