Intel-GPU

Prerequisites

Having your GPU isolated when using a VM
Passed the GPU to your Talos Machine when using a VM

Extensions for Talhelper/Clustertool

Its important to add the following Extensions to your talconfig.yaml for bootstrap:

schematic:
    customization:
        systemExtensions:
            officialExtensions:
                - siderolabs/i915
                - siderolabs/intel-ucode
                - siderolabs/mei

Adding it to your cluster

If its a fresh bootstrap you can simply follow the clustertool guide on how to bootstrap your cluster. If it is a existing cluster you will need to run clustertool talos upgrade to add the extensions to your cluster.

Adding Intel Repo for required Charts

Add the following repo to your cluster if using fluxcd:

---
# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/source.toolkit.fluxcd.io/helmrepository_v1.json
apiVersion: source.toolkit.fluxcd.io/v1
kind: HelmRepository
metadata:
  name: intel
  namespace: flux-system
spec:
  interval: 2h
  url: https://intel.github.io/helm-charts

Add intel-device-plugin-operator

Add the intel-device-plugin-operator to your cluster Example helm-release configuration:

---
# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/helm.toolkit.fluxcd.io/helmrelease_v2.json
apiVersion: helm.toolkit.fluxcd.io/v2
kind: HelmRelease
metadata:
  name: intel-device-plugin-operator
  namespace: system
spec:
  interval: 30m
  chart:
    spec:
      chart: intel-device-plugins-operator
      version: 0.32.0
      sourceRef:
        kind: HelmRepository
        name: intel
        namespace: flux-system
  install:
    crds: CreateReplace
    remediation:
      retries: 3
  upgrade:
    cleanupOnFail: true
    crds: CreateReplace
    remediation:
      strategy: rollback
      retries: 3
  dependsOn:
    - name: node-feature-discovery
      namespace: kube-system
  values:
    controllerExtraArgs: |
      - --devices=gpu

Add intel-device-plugin-gpu

Add the intel-device-plugin-gpu to your cluster Example helm-release configuration:

---
# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/helm.toolkit.fluxcd.io/helmrelease_v2.json
apiVersion: helm.toolkit.fluxcd.io/v2
kind: HelmRelease
metadata:
  name: intel-device-plugin-gpu
  namespace: system
spec:
  interval: 30m
  chart:
    spec:
      chart: intel-device-plugins-gpu
      version: 0.32.0
      sourceRef:
        kind: HelmRepository
        name: intel
        namespace: flux-system
  install:
    remediation:
      retries: 3
  upgrade:
    cleanupOnFail: true
    remediation:
      strategy: rollback
      retries: 3
  dependsOn:
    - name: intel-device-plugin-operator
      namespace: system
  values:
    name: intel-gpu-plugin
    sharedDevNum: 5
    nodeFeatureRule: true

Check if GPU is schedulable

kubectl get nodes -o=jsonpath="{range .items[*]}{.metadata.name}{'\n'}{' i915: '}{.status.allocatable.gpu\.intel\.com/i915}{'\n'}"

Example of GPU Assignment

The following shows an example on how to add the GPU to a chart. Depending on the chart you may need to adapt the workload-name.

resources:
    limits:
      gpu.intel.com/i915: 1