Install MIVisionX on Windows

Install MIVisionX on Windows#

Note

The HIP backend is not supported on Windows. The default Windows backend is OpenCL.

Prerequisites#

Windows SDK and a C++17 toolchain (Visual Studio 2019 or later)
CMake 3.10 or later
AMD drivers
OpenCL SDK (for the OpenCL backend)
OpenCV (optional — only used by RunVX for image and video display)

If OpenCV is installed, set the OpenCV_DIR environment variable to the OpenCV/build folder and add %OpenCV_DIR%\x64\vc14\bin or %OpenCV_DIR%\x64\vc15\bin to PATH.

Build#

The legacy Visual Studio .sln/.vcxproj project files have been removed. Build with CMake:

git clone https://github.com/ROCm/MIVisionX.git
cd MIVisionX

# OpenCL backend (default — requires the OpenCL SDK)
cmake -B build
cmake --build build --config Release

# CPU-only (no GPU or OpenCL SDK required)
cmake -B build -DGPU_SUPPORT=OFF
cmake --build build --config Release

You can open the CMake-generated solution in the build folder with Visual Studio, or build entirely from the command line as shown above.

Verify the build#

Use RunVX to run a sample graph:

.\build\bin\Release\runvx.exe ADD_PATH_TO\MIVisionX\samples\gdf\skintonedetect.gdf