背景
部署共享GPU集群，加速AI应用计算。
步骤
1 创建还GPU的node-group
AMI选择AL2_x86_64_GPU，自带显卡需要，省去手动安装步骤。
建议选择nvidia显卡机器。
节点打上tag: eks-node=gpu，后面安装驱动用。
2 helm安装 nvidia-device-plugin
1
2
3
4
5
6
7
8
9
|
helm repo add nvdp https://nvidia.github.io/k8s-device-plugin
helm repo update
helm upgrade -i nvdp nvdp/nvidia-device-plugin \
--namespace nvidia-device-plugin \
-f nvdia-plugin-values.yaml \
--create-namespace \
--version 0.14.5
|
nvdia-plugin-values.yaml示例，选择带显卡机器。
1
2
| nodeSelector:
eks-node: gpu
|
3 验证安装
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
| $ cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
name: gpu-pod
spec:
restartPolicy: Never
containers:
- name: cuda-container
image: nvcr.io/nvidia/k8s/cuda-sample:vectoradd-cuda10.2
resources:
limits:
nvidia.com/gpu: 1 # requesting 1 GPU
tolerations:
- key: nvidia.com/gpu
operator: Exists
effect: NoSchedule
EOF
$ k logs gpu-pod
|
参考资料