enabling-accelerators.adoc

Enabling accelerators

Before you can use an accelerator in {productname-short}, you must install the relevant software components. The installation process varies based on the accelerator type.

Prerequisites

• You have logged in to your {openshift-platform} cluster.

• You have the cluster-admin role in your {openshift-platform} cluster.

• You have installed an accelerator and confirmed that it is detected in your environment.

Procedure

1. Follow the appropriate documentation to enable your accelerator:

   • NVIDIA GPUs: See Enabling NVIDIA GPUs.

   • Intel Gaudi AI accelerators: See Enabling Intel Gaudi AI accelerators.

   • AMD GPUs: See Enabling AMD GPUs.

2. After installing your accelerator, create an accelerator profile as described in: Working with accelerator profiles.

Verification

• From the Administrator perspective, go to the Operators → Installed Operators page. Confirm that the following Operators appear:

  • The Operator for your accelerator

  • Node Feature Discovery (NFD)

  • Kernel Module Management (KMM)

• The accelerator is correctly detected a few minutes after full installation of the Node Feature Discovery (NFD) and the relevant accelerator Operator. The {openshift-platform} command line interface (CLI) displays the appropriate output for the GPU worker node. For example, here is output confirming that an NVIDIA GPU is detected:

  # Expected output when the accelerator is detected correctly
  oc describe node <node name>
  ...
  Capacity:
    cpu:                4
    ephemeral-storage:  313981932Ki
    hugepages-1Gi:      0
    hugepages-2Mi:      0
    memory:             16076568Ki
    nvidia.com/gpu:     1
    pods:               250
  Allocatable:
    cpu:                3920m
    ephemeral-storage:  288292006229
    hugepages-1Gi:      0
    hugepages-2Mi:      0
    memory:             12828440Ki
    nvidia.com/gpu:     1
    pods:               250