Intel GPU device plugin setup #

Nimbra Edge uses the Intel GPU device plugin for Kubernetes to expose Intel GPU render devices to video pods. This is required when using --transcode-accelerator quicksync on Kubernetes deployments.

Prerequisites #

• Kubernetes cluster with Intel GPU-equipped nodes
• Intel GPU kernel driver (i915) loaded on the nodes

Installation #

Deploy the Intel GPU device plugin DaemonSet:

kubectl apply -k https://github.com/intel/intel-device-plugins-for-kubernetes/deployments/nfd/overlays/node-feature-discovery
kubectl apply -k https://github.com/intel/intel-device-plugins-for-kubernetes/deployments/gpu_plugin/overlays/nfd_labeled_nodes

Alternatively, install via Helm:

helm repo add intel https://intel.github.io/helm-charts
helm install intel-gpu-plugin intel/intel-device-plugins-gpu

Verification #

Verify that nodes with Intel GPUs show gpu.intel.com/i915 as an allocatable resource:

kubectl get nodes -o json | jq '.items[].status.allocatable | select(.["gpu.intel.com/i915"])'

Expected output:

{
  "gpu.intel.com/i915": "1"
}