How to build shared object?

Hi CUDA experts,

I am trying to compile this project on my workstation. However, it shows these errors:

nvcc -gencode arch=compute_60,code=sm_60 -gencode arch=compute_70,code=sm_70 -O3 -std=c++11 --expt-extended-lambda -Xcompiler="-fopenmp" -Wreorder -lineinfo build/plan_parser.o build/execute.o -o execute

/usr/bin/ld: cannot find -lcudadevrt

/usr/bin/ld: cannot find -lcudart_static

collect2: error: ld returned 1 exit status

Makefile:33: recipe for target ‘execute’ failed

make: *** [execute] Error 1

And I found that in the directory /usr/local/cuda/lib64, there is no libcudadevrt.a. My question is: is it possible to compile those file into dynamic object .so and link them together? If yes, how to do that?