How can I tell what architecture my code is being compiled for?

Accelerated Computing CUDA CUDA Programming and Performance
1

I am writing a kernel that needs to perform a warp vote. However, if I am compiling for an architecture that does not support the warp vote (i.e. CC less than 1.2, I think) then I have a sort of ugly workaround. I would like to know if there is a way I can check whether or not the warp vote functions are defined, or otherwise check the compute capability of the target architecture of the current compile.

Thanks for reading.

2

check the macro CUDA_ARCH to see if it’s greater than or equal to 120 (page 32 in nvcc.pdf from 2.2)