container-service-for-azure-china/aks/gpu-support.md

GPU workload support on Azure China AKS

Use GPUs for compute-intensive workloads on Azure Kubernetes Service (AKS) provides detailed steps about how to run GPU workloads on AKS cluster, while there are some configurations needed to change on Azure China. e.g. following docker hub images should be changed to use dockerhub.azk8s.cn:

original image in doc	supported images on Azure China
k8s-device-plugin:1.11	dockerhub.azk8s.cn/nvidia/k8s-device-plugin:1.11
microsoft/samples-tf-mnist-demo:gpu	dockerhub.azk8s.cn/microsoft/samples-tf-mnist-demo:gpu

Below are detailed steps about how to run GPU workload on Azure China AKS cluster:

Prerequisite

Make sure your Azure subscription could create NC serial VMs, e.g. Standard_NC6s_v3, otherwise you may file a support ticket to enable that kind of VM size for your Azure subscription.

1. set up AKS cluster on GPU enabled VM

Below example sets node-vm-size as Standard_NC6s_v3 which supports GPU on Azure China, on global azure, node-vm-size could be Standard_NC6 etc.

RESOURCE_GROUP_NAME=demo-gpu1126
CLUSTER_NAME=demo-gpu1126
LOCATION=chinaeast2
az group create -n $RESOURCE_GROUP_NAME -l $LOCATION
		 
az aks create -g $RESOURCE_GROUP_NAME -n $CLUSTER_NAME --node-count 1 --node-vm-size Standard_NC6s_v3 --disable-rbac --generate-ssh-keys --kubernetes-version 1.12.6 -l $LOCATION	
az aks get-credentials -g $RESOURCE_GROUP_NAME -n $CLUSTER_NAME
kubectl get nodes

2. install GPU plugin on AKS cluster

kubectl create -f https://raw.githubusercontent.com/andyzhangx/demo/master/linux/gpu/nvidia-device-plugin-ds-mooncake.yaml

3. Run GPU workload on AKS cluster

kubectl create -f https://raw.githubusercontent.com/andyzhangx/demo/master/linux/gpu/gpu-demo-mooncake.yaml

For more detailed steps, refer to Use GPUs for compute-intensive workloads on Azure Kubernetes Service (AKS)