gcc versions error when installing OpenCV 3.1.0

[Xiaoming-Zhao]

Hi, I met a problem in installing OpenCV 3.1.0

• Host OS: Linux (Ubuntu 15.10 x86_64)
• CUDA 7.5.17

I installed CUDA first, and when I followed this installation guide, I met an error at make -j2:

[ 18%] Building NVCC (Device) object modules/core/CMakeFiles/cuda_compile.dir/src/cuda/cuda_compile_generated_gpu_mat.cu.o
In file included from /usr/local/cuda-7.5/include/cuda_runtime.h:76:0,
                 from <command-line>:0:
/usr/local/cuda-7.5/include/host_config.h:115:2: error: #error -- unsupported GNU version! gcc versions later than 4.9 are not supported!
 #error -- unsupported GNU version! gcc versions later than 4.9 are not supported!
  ^
CMake Error at cuda_compile_generated_gpu_mat.cu.o.cmake:208 (message):
  Error generating
  /home/zhaoxiaoming/opencv-3.1.0/build/modules/core/CMakeFiles/cuda_compile.dir/src/cuda/./cuda_compile_generated_gpu_mat.cu.o


modules/core/CMakeFiles/opencv_core.dir/build.make:55: recipe for target 'modules/core/CMakeFiles/cuda_compile.dir/src/cuda/cuda_compile_generated_gpu_mat.cu.o' failed
make[2]: *** [modules/core/CMakeFiles/cuda_compile.dir/src/cuda/cuda_compile_generated_gpu_mat.cu.o] Error 1
CMakeFiles/Makefile2:1184: recipe for target 'modules/core/CMakeFiles/opencv_core.dir/all' failed

About my gcc, I checked my /usr/bin/, there are three versions of gcc and g++, namely 4.8.5, 4.9.3 and 5.2.1. Because when using CUDA, it required gcc no later than 4.9, so I did the following:

sudo ln -s /usr/bin/gcc-4.9 /usr/local/cuda/bin/gcc
sudo ln -s /usr/bin/g++-4.9 /usr/local/cuda/bin/g++

And when met the OpenCV error, I did the following:

sudo ln -s /usr/bin/gcc-4.8 /usr/local/cuda/bin/gcc
sudo ln -s /usr/bin/g++-4.8 /usr/local/cuda/bin/g++

And when I usegcc --version to check, the output is 4.8.5.

I am wondering whether there is a method to let the OpenCV make -j2 procedure uses exactly the right gcc version?

Thanks!

[StevenPuttemans]

You could open up the cmake-gui and explicitly tell it which compiler you want to use from the first drop down menu I guess.

[alalek]

http://stackoverflow.com/questions/17275348/how-to-specify-new-gcc-path-for-cmake
but It is hard to say how does this help CUDA NVCC

[Rohith25]

I kinda had the same issue. Apparently NVCC needs to have its flag set manually. Add the following to cmake and it works fine

CUDA_NVCC_FLAGS="-ccbin gcc-4.9"

[bonben]

Fix by setting parameter :
//Host side compiler used by NVCC
CUDA_HOST_COMPILER:FILEPATH=/usr/bin/gcc-4.9