The function/feature is not implemented (The called functionality is disabled for current build or platform) in throw_no_cuda

, ,
1

Hi Technical Team,
Environment used for my task:

  • Windows 10
  • CUDA toolkits: 12.8
  • GPU Name : NVIDIA Geforce RTX 3070
  • OpenCV 4.10.0
  • CMake version:3.18.6

I have build OpenCV with CUDA support using CMake GUI for C++.

When I build OpenCV with CUDA support in CMake, the output is as follows:

NVIDIA CUDA:     YES (ver 12.8, CUFFT CUBLAS NVCUVID NVCUVENC)

But when I write my C++ project in Visual Studio 2019, the output I get from calling the getBuildInformation function in the main function is the following situation:

NVIDIA CUDA:      YES (ver 12.8, CUFFT CUBLAS FAST_MATH)

And in the main function, I wrote the following code:

cv::Ptr<cv::cudacodec::VideoReader> reader = cv::cudacodec::createVideoReader("video.mp4");  
if (!reader) {  
    std::cerr << "无法创建 VideoReader" << std::endl;  
    return -1;  
}

When debugging, it reports an error. The specific error messages are as follows:

OpenCV(4.10.0) Error: The function/feature is not implemented (The called functionality is disabled for current build or platform) in throw_no_cuda, file E:\OpenCV\opencv-4.10.0\opencv\sources\modules\core\include\opencv2/core/private.cuda.hpp, line 112

May I ask what causes this?

2

Either you have two builds one with and one without NVCUVID and NVCUVENC and you are linking to the one without or you didn’t build with these options and only configured CMake with them.

Either way the output from getBuildInformation confirms that you are linking against a version of OpenCV that was not built against the NVCUVID and NVCUVENC libraries.