Sorry if I'm in the wrong forum for this but it seemed reasonable. So I'm rewriting a program for CUDA, and when I compile with nvcc -O3 or nvcc -O1 everything works, but with nvcc -O0, some .cu (cuda C) files fail to link properly. Forgetting linking issues, what does the -O* flag do?
Dunno. But I did not post here just to say I didn't know. I am posting here to point you in the right direction: Ask your question to nVidia. I found this: http://forums.nvidia.com/index.php?showforum=62 .