Is it possible to compile completely using clang?

[kayaercument]

Dear all,

I'm trying to compile CUDA versions completely using clang and I'm facing some issues.

For example, original compilation sequence of level0 is below: `clang++ -DHAVE_CONFIG_H -I. -I../../../config -I/usr/include -I../../../src/cuda/common -I../../../src/common -I../../../config -MT main.o -MD -MP -MF .deps/main.Tpo -c -o main.o ./../common/main.cpp

mv -f .deps/main.Tpo .deps/main.Po /usr/bin/nvcc -gencode=arch=compute_61,code=sm_61 -I../../../src/cuda/common -I../../../src/common -I../../../config -c BusSpeedDownload.cu clang++ -L../../../src/cuda/common -L../../../src/common -o BusSpeedDownload main.o BusSpeedDownload.o -lSHOCCommon -std=c++14 -L/usr/lib/x86_64-linux-gnu/stubs -lcudadevrt -lcudart_static -lrt -lpthread -ldl -lrt -lrt

/usr/bin/nvcc -gencode=arch=compute_61,code=sm_61 -I../../../src/cuda/common -I../../../src/common -I../../../config -c BusSpeedReadback.cu clang++ -L../../../src/cuda/common -L../../../src/common -o BusSpeedReadback main.o BusSpeedReadback.o -lSHOCCommon -std=c++14 -L/usr/lib/x86_64-linux-gnu/stubs -lcudadevrt -lcudart_static -lrt -lpthread -ldl -lrt -lrt

/usr/bin/nvcc -gencode=arch=compute_61,code=sm_61 -I../../../src/cuda/common -I../../../src/common -I../../../config -c DeviceMemory.cu clang++ -L../../../src/cuda/common -L../../../src/common -o DeviceMemory main.o DeviceMemory.o -lSHOCCommon -std=c++14 -L/usr/lib/x86_64-linux-gnu/stubs -lcudadevrt -lcudart_static -lrt -lpthread -ldl -lrt -lrt

/usr/bin/nvcc -gencode=arch=compute_61,code=sm_61 -I../../../src/cuda/common -I../../../src/common -I../../../config -c MaxFlops.cu clang++ -L../../../src/cuda/common -L../../../src/common -o MaxFlops main.o MaxFlops.o -lSHOCCommon -std=c++14 -L/usr/lib/x86_64-linux-gnu/stubs -lcudadevrt -lcudart_static -lrt -lpthread -ldl -lrt -lrt`

And my attempt to compile it using clang is like this:

`clang++ MaxFlops.cu -I../../../src/cuda/common -I../../../src/common -I../../../config -L/usr/local/cuda-10.0/lib64 --cuda-gpu-arch=sm_61 --cuda-path=/usr/local/cuda-10.0 -lcudart_static -ldl -lrt -pthread -c

clang++ -L../../../src/cuda/common -L../../../src/common -o MaxFlops main.o MaxFlops.o -lSHOCCommon -std=c++14 -L/usr/lib/x86_64-linux-gnu/stubs -lcudadevrt -lcudart_static -lrt -lpthread -ldl -lrt -lrt `

I got error as: "undefined reference to `__cudaPushCallConfiguration'"

Is there any way to fix it?

Thank you for your time.