README.md

Overview

Intel AI inference end-to-end solution with RHOCP is based on the Intel® Data Center GPU Flex Series provisioning, Intel® OpenVINO™, and Red Hat OpenShift AI (RHOAI) on RHOCP. There are two AI inference modes verified with Intel® Xeon® processors and Intel Data Center GPU Flex Series with RHOCP.

• Interactive mode – RHOAI provides OpenVINO based Jupyter Notebooks for users to interactively debug the inference applications or optimize the models on RHOCP using data center GPU cards or Intel Xeon processors.
• Deployment mode – OpenVINO Model Sever (OVMS) can be used to deploy the inference workloads in data center and edge computing environments on RHOCP.

Prerequisites

• Provisioned RHOCP cluster. Follow steps here
• Provisioning Intel Data Center GPU Flex Series. Follow steps here
  • Setup node feature discovery (NFD). Follow the steps here
  • Setup out of tree drivers for Intel GPU provisioning. Follow the steps here
  • Setup Intel device plugins operator and create Intel GPU device plugin. Follow the steps here

Install RHOAI

The Red Hat certified RHOAI operator is published at Red Hat Ecosystem Catalog. You can use the command line interface (CLI) or web console to install it.

Install using CLI (To be added)

Install using web console

1. On the RHOCP web console, click Operators → OperatorHub.
2. Search RedHat OpenShift AI Operator and click Install. The operator is installed in the namespace redhat-ods-operator.

Verification

1. Navigate to Operators → Installed Operators page.
2. Ensure that in the redhat-ods-operator namespace, RedHat OpenShift AI status is InstallSucceeded
3. Click on Search -> Routes -> rhods-dashboard from the web console and access the RHOAI UI link. Note: When installing the operator, the default kfdef Custom Resource (CR) is created. This CR enables the dashboard for users to browse and launch Jupyter Notebooks projects on an RHOCP cluster. Please refer to this link for more details about kfdef.

Install OpenVINO operator

The OpenVINO operator is published at Red Hat Ecosystem Catalog. You can use the CLI or web console to install it.

Install using CLI (To be added)

Install using web console

Follow this link to install the operator via the web console.

Work with interactive mode

To enable the interactive mode, the OpenVINO notebook CR needs to be created and integrated with RHOAI.

1. Click on the create Notebook option from the web console and follow these steps to create the notebook CR.
2. Enable Intel Data Center GPU on RHOAI Dashboard- Technical Preview feature

Create AcceleratoProfile in the redhat-ods-applications namespace

$ oc apply -f https://raw.githubusercontent.com/intel/intel-technology-enabling-for-openshift/main/e2e/inference/accelerator_profile.yaml

3. Navigate to openvino-notebooks ImageStream and add the above created AcceleratorProfile key to the annotation field, as shown in the image below:

4. Navigate to Search -> Networking -> Routes from the web console and access rhods-dashboard route in the redhat-ods-applications namespace, as in the image below. Click on the location link to launch RHOAI dashboard.

5. If step 2 is successful, Intel® Data Center GPU Flex Series 140 is shown in the accelerator dropdown menu in rhods-dashboard. Users can run OpenVINO notebook image with Intel® Data Center GPU Flex Series 140 card.

Follow the link for more details on the available Jupyter Notebooks.

Work with deployment mode

1. From the web console, click on the ModelServer option in this link and follow the steps to start the OVMS instance.
2. To enable the Intel Data Center GPU, make sure to modify the OVMS instance options according to the screenshot below.

• Below images show gpu.intel.com/i915 resource requests and limits for OVMS

See Also

GPU accelerated demo with OpenVINO