NVCC Include Path

Where does NVCC look for header files to include besides the command line? I’m trying to compile some of the samples and it’s not finding the header files it needs.

john@deathStar:/usr/local/cuda-10.0/samples/1_Utilities/bandwidthTest$ make
“/usr/local/cuda-10.0”/bin/nvcc -ccbin g++ -I…/…/common/inc -m64 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_75,code=compute_75 -o bandwidthTest.o -c bandwidthTest.cu
cc1plus: fatal error: cuda_runtime.h: No such file or directory
compilation terminated.
Makefile:288: recipe for target ‘bandwidthTest.o’ failed
make: *** [bandwidthTest.o] Error 1

I can find this include file and I could just add this path to the arguments that the Makefile passes to NVCC, but I’d like to make the samples work out-of-the-box and it seems like it should be possible.

john@deathStar:/usr/local/cuda-10.0/include$ ls cuda_runtime*
cuda_runtime_api.h cuda_runtime.h

OS: Ubuntu 18.04
Nvidia Driver: 390.77
CUDA release 10.0, V10.0.130