Some problem in building the project

[Jarvisss]

Hi, nice to meet you, I am a student from China and I cloned this project and tried to build it but got into trouble. Where can I find information about how to build this project? Thanks a lot!!!

[PengfeiRen96]

你好，我编译这个工程也遇到问题了，可以一起讨论一下吗

[Jarvisss]

nice! 一起讨论吧，qq505509679

[anastasia-tkach]

Hello,
The project includes many third party libraries, first you should make sure that all the libraries are linked. For that have a look at hmodel-cpp-public/apps/ folder and build each of the "helloworld" applications. Also, you probably need to replace the line
set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS} -gencode arch=compute_50,code=sm_50")
in https://github.com/anastasia-tkach/hmodel-cpp-public/blob/master/cmake/ConfigureCUDA.cmake
by the parameters corresponding to your GPU. You can look up on the web what are those parameters.

[Jarvisss]

Thanks for your reply, I've got the third party libraries linked, but there are still many bugs in the project, such as LINE 4 #define M_PI 3.14159265358979; a ';' is useless and causes an error while compilng.
github.com/anastasia-tkach/hmodel-cpp-public/blob/master/cudax/functors/CorrespondencesFinder.h

Further, glm type dvec3 and vec3 are used together in Hmodel and cause errors like error C2664: “bool get_tangents(const glm::dvec3 &,const glm::dvec3 &,double,double,glm::dvec2 &,glm::dvec2 &,glm::dvec2 &,glm::dvec2 &)”: 无法将参数 1 从“glm::detail::tvec3<float,0>”转换为“const glm::dvec3 &” .
Have you successfully built this project?