Error building gpu module (OpenCV 2.4.13, Win7, VS 2013 and Cuda 7.5)
Hi,
I am having some issues trying to build the latest version of OpenCV (from the GIT repo). I can build OpenCV without errors if I disable WITH_CUDA in CMake, but if I enable it (I want CUDA support), then I get an error when building the gpu module. I use Windows 7, Visual Studio Community Edition 2013, and CUDA Toolkit 7.5. The CUDA samples work properly on my Visual Studio installation.
When building the gpu module, everything goes fine until it starts to build "generalized_hough.cu", and then I start getting lots of warnings. Then, I finally get an error when compiling "gftt.cu". This error says: "C:\Program Files (x86)\Windows Kits\8.1\Include\um\combaseapi.h(229): error : identifier "IUnknown" is undefined"
Below are the CMake and Visual Studio logs:
CMake: http://pastebin.com/axs4kJ5v Visual Studio build: http://pastebin.com/4yetHqxq
Anyone knows what is going wrong?
Thanks, Carlos