NVIDIA AI Enterprise

About NVIDIA AI Enterprise and Supported Platforms
Installing GPU Operator Using the vGPU Driver
- Prerequisites
- Procedure
Updating NLS Client License Token
Installing GPU Operator Using the Data Center Driver
Related Information

About NVIDIA AI Enterprise and Supported Platforms

NVIDIA AI Enterprise is an end-to-end, cloud-native suite of AI and data analytics software, optimized, certified, and supported by NVIDIA with NVIDIA-Certified Systems.

Deploying the GPU Operator with NVIDIA AI Enterprise offers two installation options.

vGPU Guest Driver

Data Center Driver

Uses a a prebuilt vGPU driver image that is only available to NVIDIA AI Enterprise customers.

It is configured to use the NVIDIA License System (NLS). Installations on virtualization platforms must use the vGPU driver installation.

Installation is performed by downloading a Bash script from NVIDIA NGC and running the script.

Uses the GPU Operator Helm chart that is publicly available and GPU driver containers that are publicly available.

You must determine the supported driver branch, such as 550, for your NVIDIA AI Enterprise release.

Installation is performed by running the helm command.

For information about supported platforms, hypervisors, and operating systems, refer to the Product Support Matrix in the NVIDIA AI Enterprise documentation.

For information about using vGPU with Red Hat OpenShift, refer to NVIDIA AI Enterprise with OpenShift.

Installing GPU Operator Using the vGPU Driver

Prerequisites

A client configuration token has been generated for the client on which the script will install the vGPU guest driver. Refer to Generating a Client Configuration Token in the NVIDIA License System User Guide for more information.
An NGC CLI API key that is used to create an image pull secret. The secret is used to pull the prebuilt vGPU driver image from NVIDIA NGC. Refer to Generating Your NGC API Key in the NVIDIA NGC Private Registry User Guide for more information.

Procedure

Export the NGC CLI API key and your email address as environment variables:

$ export NGC_API_KEY="M2Vub3QxYmgyZ..."
$ export NGC_USER_EMAIL="user@example.com"

Go to the NVIDIA GPU Operator - Deploy Installer Script web page on NVIDIA NGC.

Click the File Browser tab, identify your NVIDIA AI Enterprise release, click , and select Download File.

Copy the downloaded script to the same directory as the client configuration token.
Rename the client configuration token that you downloaded to client_configuration_token.tok. Originally, the client configuration token is named to match the pattern: client_configuration_token_mm-dd-yyyy-hh-mm-ss.tok.
From the directory that contains the downloaded script and the client configuration token, run the script:
```
$ bash gpu-operator-nvaie.sh install
```

Updating NLS Client License Token

In case the NLS client license token needs to be updated, please use the following procedure:

Create an empty vGPU license configuration file:

$ sudo touch gridd.conf

Generate and download a new NLS client license token. Please refer to Section 4.6 of the NLS User Guide for instructions.

Rename the NLS client license token that you downloaded to client_configuration_token.tok.

Create a new licensing-config-new ConfigMap object in the gpu-operator namespace (make sure the name of the configmap is not already used in the kubernetes cluster). Both the vGPU license configuration file and the NLS client license token will be added to this ConfigMap:

$ kubectl create configmap licensing-config-new \
    -n gpu-operator --from-file=gridd.conf --from-file=<path>/client_configuration_token.tok

Edit the clusterpolicies by using the command:

$ kubectl edit clusterpolicies.nvidia.com

Go to the driver section and replace the following argument:

licensingConfig:
    configMapName: licensing-config

with

licensingConfig:
    configMapName: licensing-config-new

Write and exit from the kubectl edit session (you can use :qw for instance if vi utility is used)

GPU Operator sequentially redeploys all the driver pods with this new licensing information.

Installing GPU Operator Using the Data Center Driver

This installation method is available for bare metal clusters or any cluster that does not use virtualization.

You must install the driver that matches the supported driver branch for your NVIDIA AI Enterprise release. The following list summarizes the driver branches for each release.

v5.x: 550 branch
v4.x: 535 branch
v3.x: 525 branch
v1.x: 470 branch

For newer releases, you can confirm the the supported driver branch by performing the following steps:

Refer to the release documentation for NVIDIA AI Enterprise and access the documentation for your release.
In the release notes, identify the supported NVIDIA Data Center GPU Driver branch.

For example, the Supported Hardware and Software for the 5.1 release indicates that the release uses the 550.90.07 version of the Linux driver.
Refer to GPU Operator Component Matrix to identify the recommended driver version that uses the same driver branch, 550, in this case.

After identifying the correct driver version, refer to Installing the NVIDIA GPU Operator to install the Operator by using Helm.