I am trying to run one of the cuda_sample files (V.11) by using CMake. However, GPU breakpoints are not working for me with CMake 3.18/VS2017/Nsight 2020.1.

In my CMakeLists.txt, I use set(CUDA_NVCC_FLAGS_DEBUG “-g -G”). Is there something else that might be missed? Or should I use somthing else?

You can also see the following output message:

[CUDA] Module loaded. 0000026da0d4ee50 Debug_Line: false Debug_Info: false PTX_Line: false PTX_Reg: false PTX_Type: false

Thanks in advance for your help.