Has anyone built OpenCV with CUDA compute 7.0?
I'm currently attempting to build OCV with CUDA compute level 7.0 enabled using visual studio 2017, and I keep getting an error
I see that in the latest cuda cmake file, Volta/7.0 have been added, but every time I try to run a build I get:
Error LNK2019 unresolved external symbol __cudaRegisterLinkedBinary_54_tmpxft_00003c60_00000000_14_gpu_mat_compute_70_cpp1_ii_71482d89 referenced in function "void __cdecl __sti____cudaRegisterAll(void)" (?__sti____cudaRegisterAll@@YAXXZ) opencv_core C:\OCV40\modules\core\cuda_compile_generated_gpu_mat.cu.obj
It seems to be unhappy about CUDA 70, so is it expected that it is supported at the moment, or is this expected behavior? Thanks
What GPU device are you running? Could you give more specifics? We have 8 V100 cards in a DGX-1 setup here and OpenCV compiles just fine with compute capability 7.0.
Ah damn, thanks for that additional datapoint. So it's at least not fundamental to opencv in general, so I wonder if it's related to my specific environment. (And the intention is to be using it with Quadro GV100's, but since it won't even build for me I haven't even been able to involve the GPU yet.)
My build setup is Visual Studio 2017 + CUDA 9.2 on Windows 10 64bit. I'm assuming you're setup with Linux since you're on a DGX-1? I suppose it could be unique to VS + Windows possibly.
Yeah its a linux system, VS + windows on a DGX1 is a no go :D