Kria™ KV260 Vision AI Starter Kit Tutorial

Integrating the overlay into the Platform

Integrate Accelerator Overlay

Introduction

This document provides the steps to integrate the overlay consisting of accelerated functions into the platform.

Prerequisites

  • AMD Vitis™ Unified Software Platform 2021.1

Accessing the Tutorial Reference Files

NOTE: Skip the following steps if the design files have already been cloned and extracted to a working repository.

  1. To access the reference files, type the following into a terminal:

    git clone --branch xlnx_rel_v2022.1 --recursive https://github.com/Xilinx/kria-vitis-platforms.git

  2. Navigate to the kria-vitis-platforms/kv260-vitis which is the working directory.

Integrating an Overlay (Accelerator Functions) into a Vitis Extensible Platform

  1. Go to the working directory.

    cd $working_dir/

  2. To compile and integrate the overlay into the platform, run the following command. The Makefile calls a lower-level Makefile to run Vitis. If a platform is not already available, it builds that as well.

    make overlay OVERLAY=<application_name>

    Applications currently supported are listed as follows:

Application name Platform used
smartcam kv260_ispMipiRx_vcu_DP
aibox-reid kv260_vcuDecode_vmixDP
defect-detect kv260_ispMipiRx_vmixDP
nlp-smartvision kv260_ispMipiRx_rpiMipiRx_DP

  1. The generated bitfile and xclbin will be located at:

    $working_dir/overlays/examples/<application_name>/binary_container_1/link/int/system.bit   
$working_dir/overlays/examples/<application_name>/binary_container_1/*.xclbin

    The xclbin and bitfile generation might take couple of hours depending on the system specification. They will be used to build the PetaLiunx image.

References

For more information on how to integrate an accelerator into a platform, refer to the Vitis Unified Software Platform Documentation: Application Acceleration Development (UG1393).