How to compile and execute fortran code with CUDA 8.0?

Question

I have a NVIDIA Tesla C2075 GPU card and CUDA 8.0 toolkit.

./deviceQuery and ./bandwidthTest is also passed. Using "nvcc" I could pass