ROCm Performance Primitives Library

AMD ROCm Performance Primitives (RPP) library is a comprehensive high-performance computer vision library for AMD processors with HIP/OpenCL/CPU back-ends.

Latest Release

Top level design

Supported Functionalities and Variants

Supported Functionalities List

Supported Functionalities Samples

Documentation

Run the steps below to build documentation locally.

cd docs

pip3 install -r .sphinx/requirements.txt

python3 -m sphinx -T -E -b html -d _build/doctrees -D language=en . _build/html

Prerequisites

• OS

  • Ubuntu 20.04/22.04
  • CentOS 7/8
  • RHEL 8/9
  • SLES - 15-SP3

• ROCm supported hardware

• ROCm 5.4.3 and above

• Clang Version 5.0.1 and above

  • Ubuntu 20/22

    sudo apt-get install clang
    

  • CentOS 7

    sudo yum install llvm-toolset-7-clang llvm-toolset-7-clang-analyzer llvm-toolset-7-clang-tools-extra
    scl enable llvm-toolset-7 bash
    

  • CentOS 8 and RHEL 8/9

    sudo yum install clang
    

  • SLES 15-SP3

    sudo zypper install llvm-clang
    

• CMake Version 3.5 and above

• Boost Version 1.72 and above

  wget https://boostorg.jfrog.io/artifactory/main/release/1.72.0/source/boost_1_72_0.tar.gz
  tar -xzvf boost_1_72_0.tar.gz
  cd boost_1_72_0
  ./bootstrap.sh
  ./b2 install
  

  • NOTE: Install from source

• IEEE 754-based half-precision floating-point library - half.hpp

  wget https://sourceforge.net/projects/half/files/half/1.12.0/half-1.12.0.zip
  unzip half-1.12.0.zip -d half-files
  sudo mkdir /usr/local/include/half
  sudo cp half-files/include/half.hpp /usr/local/include/half
  

Prerequisites for Test Suite

• OpenCV 3.4.0/4.5.5 - pre-requisites

  sudo apt-get update
  sudo -S apt-get -y --allow-unauthenticated install build-essential libgtk2.0-dev libavcodec-dev libavformat-dev libswscale-dev python-dev python-numpy
  sudo -S apt-get -y --allow-unauthenticated install libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libdc1394-22-dev unzip wget
  

• OpenCV 3.4.0 /4.5.5 - download

  wget https://github.com/opencv/opencv/archive/3.4.0.zip
  unzip 3.4.0.zip
  cd opencv-3.4.0/
  

  OR

  wget https://github.com/opencv/opencv/archive/4.5.5.zip
  unzip 4.5.5.zip
  cd opencv-4.5.5/
  

• OpenCV 3.4.0/4.5.5 - installation

  mkdir build
  cd build
  cmake -D WITH_GTK=ON -D WITH_JPEG=ON -D BUILD_JPEG=ON -D WITH_OPENCL=OFF -D WITH_OPENCLAMDFFT=OFF -D WITH_OPENCLAMDBLAS=OFF -D WITH_VA_INTEL=OFF -D WITH_OPENCL_SVM=OFF -D CMAKE_INSTALL_PREFIX=/usr/local ..
  sudo -S make -j128 <Or other number of threads to use>
  sudo -S make install
  sudo -S ldconfig
  

• TurboJpeg installation

  sudo apt-get install nasm
  sudo apt-get install wget
  git clone -b 2.0.6.1 https://github.com/rrawther/libjpeg-turbo.git
  cd libjpeg-turbo
  mkdir build
  cd build
  cmake -DCMAKE_INSTALL_PREFIX=/usr \
        -DCMAKE_BUILD_TYPE=RELEASE  \
        -DENABLE_STATIC=FALSE       \
        -DCMAKE_INSTALL_DOCDIR=/usr/share/doc/libjpeg-turbo-2.0.3 \
        -DCMAKE_INSTALL_DEFAULT_LIBDIR=lib  \
        ..
  make -j$nproc
  sudo make install
  

Build & Install RPP

The ROCm Performance Primitives (RPP) library has support for three backends: HIP, OpenCL, and CPU:

• Building RPP with the HIP backend (default):

$ git clone https://github.com/GPUOpen-ProfessionalCompute-Libraries/rpp.git
$ mkdir build && cd build
$ cmake -DBACKEND=HIP ../rpp
$ make -j8
$ sudo make install

• Building RPP with the OPENCL backend:

$ git clone https://github.com/GPUOpen-ProfessionalCompute-Libraries/rpp.git
$ mkdir build && cd build
$ cmake -DBACKEND=OCL ../rpp
$ make -j8
$ sudo make install

• Building RPP with the CPU backend:

$ git clone https://github.com/GPUOpen-ProfessionalCompute-Libraries/rpp.git
$ mkdir build && cd build
$ cmake -DBACKEND=CPU ../rpp
$ make -j8
$ sudo make install

Test Functionalities

CPU installation

$ cd rpp/utilities/rpp-unittests/HOST_NEW
$ ./testAllScript.sh

OCL installation

$ cd rpp/utilities/rpp-unittests/OCL_NEW
$ ./testAllScript.sh

HIP installation

$ cd rpp/utilities/rpp-unittests/HIP_NEW
$ ./testAllScript.sh

MIVisionX Support - OpenVX Extension

MIVisionX RPP Extension vx_rpp supports RPP functionality through OpenVX Framework.

Technical Support

Please email [email protected] for questions, and feedback on AMD RPP.

Please submit your feature requests, and bug reports on the GitHub issues page.

Release Notes

Latest Release

Changelog

Review all notable changes with the latest release

Tested configurations

• Linux distribution
  • Ubuntu - 20.04 / 22.04
  • CentOS - 7 / 8
  • RedHat - 8 / 9
  • SLES - 15-SP3
• ROCm: rocm-core - 5.4.0.50400-72
• Protobuf - V3.12.4
• OpenCV - 4.6.0