I am trying to install OpenCV 3.2 on a Linux machine (Arch Linux) that has CUDA 10. I cannot remove CUDA 10 because I have other applications (Tensorflow, OpenCV 4, etc) that require CUDA 10. I need OpenCV 3.2 for an older machine learning application I am trying to run however, I am running into errors when building OpenCV 3.2. One of the errors I am dealing with is:
Recipe for target cuda_compile_generated_gpu_mat.cu.o failed
I have already tried disabling dnn_modern when running make.
Note I have also tried changing lines in the make file as suggested by this post: CMake Error: Variables are set to NOTFOUND
The output of the make error after typing make -j8
is:
1 error detected in the compilation of "/tmp/tmpxft_000033aa_00000000-12_gpu_mat.compute_61.cpp1.ii".
CMake Error at cuda_compile_generated_gpu_mat.cu.o.cmake:266 (message):
Error generating file
/home/patelvrajn/Development/opencv-3.2.0/build/modules/core/CMakeFiles/cuda_compile.dir/src/cuda/./cuda_compile_generated_gpu_mat.cu.o
make[2]: *** [modules/core/CMakeFiles/opencv_core.dir/build.make:65: modules/core/CMakeFiles/cuda_compile.dir/src/cuda/cuda_compile_generated_gpu_mat.cu.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1346: modules/core/CMakeFiles/opencv_core.dir/all] Error 2
make: *** [Makefile:163: all] Error 2
***OpenCV 3.2 and 4.0 are needed for both Python and C++ in my case.
Thanks in advance for any help!