Compile the cuda6.0 routine of Jetson TK1 with OpenGL

First, prepare the file system downloaded from the NVIDIA website and decompress it. Copy the <decompress path>/usr/lib/ARM-Linux-gnueabihf folder in the file system to the root directory of the host and maintain the relative directory structure, copy the <decompressed path>/lib/ARM-Linux-gnueabihf folder in the file system to the root directory of the host and maintain the relative directory structure. Use nsight to open the routines that come with the cuda-toolkit-6-0, change the target platform to arm in the Project Settings, change the x86_64 words in the path to armv7l, add the/usr/lib/ARM-Linux-gnueabihf and/lib/ARM-Linux-gnueabihf paths to the path of the dynamic library to use the nsight compilation function. If the libglu. So file is missing during compilation, you can create a symbolic link with the same name in the above path to recompile the library file such as libglu. so.1.

In addition, you can add or modify the above path in makefile and compile it using the make armv7 = 1 command.

Zookeeper

Compile the cuda6.0 routine of Jetson TK1 with OpenGL