Can this be used with Nvidia-docker?

can this be used with Nvidia-docker ?

I’ve noticed containers are just deployed using docker, and not nvidia-docker which activates all the GPUs in the server! So the container is useless for GPU related workloads.

There is a feature request for this on our GitHub repo, you can track this here

I saw the Feature Request ?

will that work with nvidia-docker on a Nvidia DGX-1.

As --runtime and ‘–gpus’ is not a syntax which works, with the docker version we have.

Okay, it would appear, we would need to wait for a supported update from Nvidia for the DGX-1, and then we can use the later version of Docker.