Skip to main content

Nvidia Gpus

L40s GPU Node​

The L40s GPU node is configured as a worker node within the cluster.

Node Information​

  • The GPU node is fully integrated into the Kubernetes cluster
  • GPU resources are available for scheduling container workloads
  • Supports NVIDIA GPU workloads with proper resource allocation

Example Container Workload​

Here's an example deployment that utilizes the GPU node:

yaml
1234567891011121314
apiVersion: v1
kind: Pod
metadata:
  name: gpu-test
spec:
  containers:
  - name: cuda-container
    image: nvidia/cuda:12.0.0-base-ubuntu22.04
    command: ["nvidia-smi"]
    resources:
      limits:
        nvidia.com/gpu: 1
  nodeSelector:
    nvidia.com/gpu.product: NVIDIA-L40S

To deploy and test:

bash
12
kubectl apply -f gpu-test.yaml
kubectl logs gpu-test
tip

Make sure to include GPU resource requests in your pod specifications to ensure proper scheduling on the L40s node.

Last updated on