nvcc include search path

I’m trying to compile NVIDIA_CUDA-10.1_Samples/1_Utilities/deviceQuery on a new machine but failed:

$ make
deviceQuery.cpp:16:10: fatal error: cuda_runtime.h: No such file or directory
 #include <cuda_runtime.h>

          ^~~~~~~~~~~~~~~~

Then I tried a dry run:

$ nvcc deviceQuery.cpp --dryrun
#$ _SPACE_=
#$ _CUDART_=cudart
#$ _HERE_=/usr/local/cuda-10.1/bin
#$ _THERE_=/usr/local/cuda-10.1/bin
#$ _TARGET_SIZE_=
#$ _TARGET_DIR_=
#$ _TARGET_DIR_=targets/x86_64-linux
#$ TOP=/usr/local/cuda-10.1/bin/..
#$ NVVMIR_LIBRARY_DIR=/usr/local/cuda-10.1/bin/../nvvm/libdevice
#$ LD_LIBRARY_PATH=/usr/local/cuda-10.1/bin/../lib:/usr/local/cuda-10.1/lib64:/usr/local/cuda-10.1/lib64:/usr/local/cuda-9.0/lib64:/usr/local/cuda/lib64
#$ PATH=/usr/local/cuda-10.1/bin/../nvvm/bin:/usr/local/cuda-10.1/bin:/usr/local/cuda-10.1/bin:/home/user/bin:/usr/local/bin:/usr/local/cuda/bin:/usr/local/cuda-10.1/bin:/usr/local/cuda-9.0/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games
#$ INCLUDES="-I/usr/local/cuda-10.1/bin/../targets/x86_64-linux/include"
#$ LIBRARIES=  "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib/stubs" "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib"
#$ CUDAFE_FLAGS=
#$ PTXAS_FLAGS=
#$ gcc -std=c++14 -c -x c++ -D__NVCC__  "-I/usr/local/cuda-10.1/bin/../targets/x86_64-linux/include"    -D__CUDACC_VER_MAJOR__=10 -D__CUDACC_VER_MINOR__=1 -D__CUDACC_VER_BUILD__=105 -m64 -o "/tmp/tmpxft_0000241a_00000000-4_deviceQuery.o" "deviceQuery.cpp"
#$ nvlink --arch=sm_30 --register-link-binaries="/tmp/tmpxft_0000241a_00000000-2_a_dlink.reg.c"  -m64   "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib/stubs" "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib" -cpu-arch=X86_64 "/tmp/tmpxft_0000241a_00000000-4_deviceQuery.o"  -o "/tmp/tmpxft_0000241a_00000000-5_a_dlink.sm_30.cubin"
#$ fatbinary --create="/tmp/tmpxft_0000241a_00000000-3_a_dlink.fatbin" -64 -link "--image=profile=sm_30,file=/tmp/tmpxft_0000241a_00000000-5_a_dlink.sm_30.cubin" --embedded-fatbin="/tmp/tmpxft_0000241a_00000000-3_a_dlink.fatbin.c"
#$ rm /tmp/tmpxft_0000241a_00000000-3_a_dlink.fatbin
#$ gcc -std=c++14 -c -x c++ -DFATBINFILE="\"/tmp/tmpxft_0000241a_00000000-3_a_dlink.fatbin.c\"" -DREGISTERLINKBINARYFILE="\"/tmp/tmpxft_0000241a_00000000-2_a_dlink.reg.c\"" -I. -D__NV_EXTRA_INITIALIZATION= -D__NV_EXTRA_FINALIZATION= -D__CUDA_INCLUDE_COMPILER_INTERNAL_HEADERS__  "-I/usr/local/cuda-10.1/bin/../targets/x86_64-linux/include"    -D__CUDACC_VER_MAJOR__=10 -D__CUDACC_VER_MINOR__=1 -D__CUDACC_VER_BUILD__=105 -m64 -o "/tmp/tmpxft_0000241a_00000000-6_a_dlink.o" "/usr/local/cuda-10.1/bin/crt/link.stub"
#$ g++ -m64 -o "a.out" -std=c++14 -Wl,--start-group "/tmp/tmpxft_0000241a_00000000-6_a_dlink.o" "/tmp/tmpxft_0000241a_00000000-4_deviceQuery.o"   "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib/stubs" "-L/usr/local/cuda-10.1/bin/../targets/x86_64-linux/lib" -lcudadevrt  -lcudart_static  -lrt -lpthread  -ldl  -Wl,--end-group

I did find cuda_runtime.h in the include folder:

$ ls /usr/local/cuda-10.1/bin/../targets/x86_64-linux/include | grep cuda_runtime.h
cuda_runtime.h

I also set PATH and LD_LIBRARY_PATH:

export PATH=/usr/local/cuda-10.1/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Don’t know what’s going on here. Is the default include search path determined when compiling CUDA? How can I tell the nvcc default include search path?