Skip to main content

AMD Now Shipping Accelerated, Conformant OpenVX 1.3 in Open Source

AMD’s Radeon MIVisionX computer vision and machine intelligence toolkit is now delivering a conformant highly optimized OpenVX 1.3 vision function library. The OpenVX implementation supports both CPU execution and GPU acceleration through OpenCL or HIP backends. This OpenVX 1.3 implementation targets a wide range of computer vision applications, including small embedded x86 CPUs, APUs, discrete GPUs, and heterogeneous servers. The implementation is fully open-sourced on GitHub at https://github.com/GPUOpen-ProfessionalCompute-Libraries/MIVisionX.