nvcc fatal : A single input file is required for a non-link phase when an outputfile is specified

[xmicroshell]

I'm a beginner about this,and when I was compile the gemm_op.so library,I was caught in a problem.
when i run make in the directory,there is a mistake as:
nvcc -std=c++11 -c -o obj/xnor_gemm_kernel.cu.o src/xnor_gemm_kernel.cu.cc -I -D GOOGLE_CUDA=1 -x cu -Xcompiler -fPIC --expt-relaxed-constexpr --Wno-deprecated-gpu-targets nvcc fatal : A single input file is required for a non-link phase when an outputfile is specified Makefile:2: recipe for target 'all' failed make: *** [all] Error 1
I wonder if its because my win system.But on my friends' linux computer it's also unworkable.
My environment is:

• windows10
• gcc version 6.3.0 (MinGW.org GCC-6.3.0-1)
• python 3.6 & python 3.5
• tensorflow 1.2.0
• Cuda compilation tools, release 8.0, V8.0.60

Thanks for anyone who can give me some advices

[kuanzi]

you should set the variable ${CUDA_ROOT}, it works for me. @xmicroshell