... | @@ -201,7 +201,7 @@ Starting from CUDA 10.0, the CUDA images require using nvidia-docker v2 and won' |
... | @@ -201,7 +201,7 @@ Starting from CUDA 10.0, the CUDA images require using nvidia-docker v2 and won' |
|
Not currently, support for Swarmkit is still being worked on in the upstream Moby project. You can track our progress [here](https://github.com/moby/moby/issues/33439).
|
|
Not currently, support for Swarmkit is still being worked on in the upstream Moby project. You can track our progress [here](https://github.com/moby/moby/issues/33439).
|
|
|
|
|
|
#### Do you support Docker Compose?
|
|
#### Do you support Docker Compose?
|
|
Yes, use Compose format `2.3` and add `runtime: nvidia` to your GPU service. Docker Compose must be version [1.19.0](https://github.com/docker/compose/releases/tag/1.19.0) or higher. You can find an example [here](https://github.com/NVIDIA/gpu-monitoring-tools/blob/master/exporters/prometheus-dcgm/docker-compose.yml).
|
|
Yes, use Compose format `2.3` and add `runtime: nvidia` to your GPU service. Docker Compose must be version [1.19.0](https://github.com/docker/compose/releases/tag/1.19.0) or higher. You can find an example [here](https://github.com/NVIDIA/gpu-monitoring-tools/master/exporters/prometheus-dcgm/docker/docker-compose.yml).
|
|
|
|
|
|
#### Do you support Kubernetes?
|
|
#### Do you support Kubernetes?
|
|
Since Kubernetes 1.8, the recommended way is to use our official [device plugin](https://github.com/NVIDIA/k8s-device-plugin).
|
|
Since Kubernetes 1.8, the recommended way is to use our official [device plugin](https://github.com/NVIDIA/k8s-device-plugin).
|
... | | ... | |