GPU Device Plugin Configuration

Intel GPU plugin enables Kubernetes clusters to utilize Intel GPUs for hardware acceleration of various workloads.

This section describes how to enable and use the Intel GPU device plugin in StarlingX.

Prerequisites

  • The host should have Intel GPU hardware. For supported GPU devices, refer to Intel GPU plugin documentation for more details: Intel GPU device plugin for Kubernetes.

  • Node Feature Discovery application must be installed using the following commands:

    ~(keystone_admin)]$ system application-upload /usr/local/share/applications/helm/node-feature-discovery-<version>.tgz
    ~(keystone_admin)]$ system application-apply node-feature-discovery
    

    Replace <version> with the latest version number.

Enable Intel GPU Device Plugin

  1. Locate the application tarball in the /usr/local/share/applications/helm directory. For example:

    /usr/local/share/applications/helm/intel-device-plugins-operator-<version>.tgz

  2. Upload the application using the following command.

    ~(keystone_admin)]$ system application-upload intel-device-plugins-operator-<version>.tgz
    

    Replace <version> with the latest version number.

  3. Verify that the application has been uploaded successfully.

    ~(keystone_admin)]$ system application-list
    
  4. Check the helm chart status using the following command:

    ~(keystone_admin)]$ system helm-override-list intel-device-plugins-operator --long
    
  5. Enable GPU Helm chart using the following command:

    ~(keystone_admin)]$ system helm-chart-attribute-modify --enabled true intel-device-plugins-operator intel-device-plugins-gpu intel-device-plugins-operator
    
  6. Apply the application using the following command:

    ~(keystone_admin)]$ system application-apply intel-device-plugins-operator
    
  7. Monitor the status of the application using one of the following commands.

    ~(keystone_admin)]$ watch -n 5 system application-list
    

    OR

    ~(keystone_admin)]$ watch kubectl get pods -n intel-device-plugins-operator
    
  8. Pods can be checked using the following command:

    $ kubectl get pods -n intel-device-plugins-operator
    

Use Intel GPU Device Plugin

For information related to using GPU device plugin, see Testing and Demos.