ROCm/rpp

[!NOTE] The published documentation is available at ROCm Performance Primitives (RPP) in an organized, easy-to-read format, with search and a table of contents. The documentation source files reside in the docs folder of this repository. As with all ROCm projects, the documentation is open source. For more information on contributing to the documentation, see Contribute to ROCm documentation.

AMD ROCm Performance Primitives (RPP) library is a comprehensive, high-performance computer vision library for AMD processors that have HIP, or CPU backends.

Latest release

Supported Augmentations / Primitives

RPP supports various 2D image, 3D image (voxel), audio and miscellaneous augmentations and primitives as listed below.

Supported 2D Image Augmentations Samples

Supported 3D Image Augmentations Samples

slice (3D slice - 100x200 from 240x240x155):

Supported Audio Augmentations Samples

Spectrogram functionality output represented as an image:

Prerequisites

Operating Systems

• Linux
  • Ubuntu - 22.04 / 24.04
  • RedHat - 8 / 9
  • SLES - 15 SP7

Hardware

• CPU: AMD64
• GPU: AMD Radeon™ Graphics / AMD Instinct™ Accelerators

[!IMPORTANT]

• ROCm-supported hardware required for HIP backend
• gfx908 or higher GPU required

• Install ROCm 7.0.0 or later with amdgpu-install: Required usecase:rocm

[!IMPORTANT] sudo amdgpu-install --usecase=rocm

Compiler

• AMD Clang++ Version 18.0.0 or later - installed with ROCm

[!NOTE]

• For CPU only backend use Clang Version 5.0.1 or later

   sudo apt install clang
  

• To use GNU compiler or custom compilers use -D CMAKE_CXX_COMPILER during build

Libraries

• CMake Version 3.10 or later

  sudo apt install cmake
  

• HIP

  sudo apt install hip-dev
  

• OpenMP

  sudo apt install openmp-extras-dev
  

• Half-precision floating-point library - Version 1.12.0 or later

  sudo apt install half
  

[!IMPORTANT]

• Required compiler support
  • C++17
  • OpenMP
  • Threads
• On Ubuntu 22.04 - Additional package required: libstdc++-12-dev

sudo apt install libstdc++-12-dev

[!NOTE]

• All package installs are shown with the apt package manager. Use the appropriate package manager for your operating system.

Installation instructions

The installation process uses the following steps:

• ROCm-supported hardware install verification

• Install ROCm 7.0.0 or later with amdgpu-install with --usecase=rocm

[!IMPORTANT] Use either package install or source install as described below.

Package install

Install RPP runtime, development, and test packages.

• Runtime package - rpp only provides the rpp library librpp.so
• Development package - rpp-dev/rpp-devel provides the library, header files, and samples
• Test package - rpp-test provides CTest to verify installation

[!NOTE] Package install will auto install all dependencies.

Ubuntu

sudo apt install rpp rpp-dev rpp-test

RHEL

sudo yum install rpp rpp-devel rpp-test

SLES

sudo zypper install rpp rpp-devel rpp-test

Source build and install

• Clone RPP git repository

  git clone https://github.com/ROCm/rpp.git
  

HIP Backend

mkdir build-hip
cd build-hip
cmake ../rpp
make -j8
sudo make install

Running Tests

After installing RPP, refer to the Verify installation section below for instructions on running tests.

Verify installation

The installer will copy

• Libraries into ${ROCM_PATH}/lib
• Header files into ${ROCM_PATH}/include/rpp
• Samples, and test folder into ${ROCM_PATH}/share/rpp
• Documents folder into ${ROCM_PATH}/share/doc/rpp

Verify with rpp-test package

Test package will install CTest module to test rpp. Follow below steps to test package install

mkdir rpp-test && cd rpp-test
cmake ${ROCM_PATH}/share/rpp/test/
ctest -VV

[!NOTE]

• Ubuntu: Install Nifti-Imaging to run all tests

git clone https://github.com/NIFTI-Imaging/nifti_clib.git
cd nifti_clib
git reset --hard 84e323cc3cbb749b6a3eeef861894e444cf7d788
mkdir build && cd build && cmake ..
sudo make -j$nproc install

• SLES/RHEL: Install prerequisites to run all tests

Test Functionalities

To test latest Image/Voxel/Audio/Miscellaneous functionalities of RPP using a python script please view AMD ROCm Performance Primitives (RPP) Test Suite

Adding RPP to your CMake project

To add RPP to your CMake project, you can use the following code after installation:

find_package(rpp REQUIRED)
target_link_libraries(your_target PRIVATE rpp::rpp)

HIP backend support is automatic: rpp::rpp transitively propagates the HIP include paths and link libraries, and rpp/rpp.h includes rpp_backend.h which sets RPP_BACKEND_HIP for your compiled sources.

[!NOTE] find_package(rpp REQUIRED) sets the following variables in your CMake project:

• rpp_BACKEND_TYPE - "HIP" or "CPU" — useful for conditional CMake logic (e.g. adding HIP-specific sources)
• rpp_AUDIO_AUGMENTATIONS_SUPPORT - ON or OFF

[!TIP] If CMake is unable to find RPP, the following fixes can be tried:

• Ensure ${ROCM_PATH}/bin is in your PATH: export PATH=${ROCM_PATH}/bin:$PATH.
• Ensure CMAKE_PREFIX_PATH includes ${ROCM_PATH}/lib/cmake.

MIVisionX support - OpenVX extension

MIVisionX RPP extension vx_rpp supports RPP functionality through the OpenVX Framework.

Technical support

For RPP questions and feedback, you can contact us at [email protected].

To submit feature requests and bug reports, use our GitHub issues page.

Documentation

You can build our documentation locally using the following code:

• Sphinx

  cd docs
  pip3 install -r .sphinx/requirements.txt
  python3 -m sphinx -T -E -b html -d _build/doctrees -D language=en . _build/html
  

• Doxygen

  doxygen .Doxyfile
  

Release notes

All notable changes for each release are added to our changelog.

Tested configurations

• Linux distribution
  • Ubuntu - 22.04 / 24.04
  • RedHat - 8 / 9
  • SLES - 15 SP7
• ROCm: rocm-core - 7.0.0+
• CMake - Version 3.10+
• AMD Clang++ - Version 18.0.0+
• Half - IEEE 754-based half-precision floating-point library - Version 1.12.0 / package V1.12.0
• OpenCV - 4.6.0