Creating CUDA C library using CMake

Question

I've the following code structure for my library

|-build
|-example
    |-CMakeLists.txt
    |-main_VecAdd.cu
|-include
    |-VonNeumann
        |-utils.h
        |-VecAdd.h
    |-CMakeLists.txt
|-src
    |-CMakeLists.txt
    |-utils.c
    |-VecAdd.c
|-CMakeLists.txt

CMakeLists.txt for folders example, include, src and main_folder are as follows (in order respectively):

project(VonNeumann LANGUAGES CUDA C)

add_executable(main_VecAdd main_VecAdd.cu)

set_target_properties(VonNeumann PROPERTIES CUDA_SEPARABLE_COMPILATION ON)

target_link_libraries(main_VecAdd 
    PRIVATE
    VonNeumann)

project(VonNeumann LANGUAGES CUDA C)

project(VonNeumann LANGUAGES CUDA C)

add_library(VonNeumann utils.c VecAdd.c)

target_include_directories(VonNeumann
    PUBLIC
    ${CMAKE_CURRENT_SOURCE_DIR}/../include)

cmake_minimum_required(VERSION 3.16.3)
project(VonNeumann LANGUAGES CUDA C)

# Adding subdirectories
add_subdirectory(include)
add_subdirectory(src)

# Giving an option to include code in examples or not
option(BUILD_EXAMPLES "Whether or not to build examples" ON)
if (BUILD_EXAMPLES)
    message(STATUS "Building examples...")
    add_subdirectory(example)
endif()

When I run cmake .. from build directory, everything works well however I get the following error when I run make command:

[ 16%] Linking C static library libVonNeumann.a
[ 50%] Built target VonNeumann
[ 66%] Linking CUDA device code CMakeFiles/main_VecAdd.dir/cmake_device_link.o
[ 83%] Linking CUDA executable main_VecAdd
/usr/bin/ld: CMakeFiles/main_VecAdd.dir/main_VecAdd.cu.o: in function `main':
tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x111): undefined reference to `RandVecFP32(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x122): undefined reference to `RandVecFP32(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x133): undefined reference to `ZeroVecFP32(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x144): undefined reference to `PrintVec(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x155): undefined reference to `PrintVec(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x166): undefined reference to `PrintVec(float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x17d): undefined reference to `VecAddFP32(float*, float*, float*, int)'
/usr/bin/ld: tmpxft_000058c9_00000000-6_main_VecAdd.cudafe1.cpp:(.text+0x19a): undefined reference to `PrintVec(float*, int)'
collect2: error: ld returned 1 exit status
make[2]: *** [example/CMakeFiles/main_VecAdd.dir/build.make:105: example/main_VecAdd] Error 1
make[1]: *** [CMakeFiles/Makefile2:163: example/CMakeFiles/main_VecAdd.dir/all] Error 2
make: *** [Makefile:84: all] Error 2

I know that the error is because of failed linking but I can't figure out how to resolve this.

P.S. I've tested this same code without CUDA part and it works just fine.

No, but then you would need to modify the CUDA code to export the symbols you need to call from your C code with C linkage — talonmies, Oct 16 '22 at 00:28
As I'm just starting out, I think it's easier to switch my code to C++. Thanks a lot. — Tushar Gautam, Oct 16 '22 at 00:30

einpoklum · Accepted Answer · 2022-10-16T10:18:29.553

As @talonmies suggests, the problem is likely you trying to use functions implemented in your .cu file, in your .c file: CUDA uses "C++ linkage", that the function symbols appear in a certain way in the compiled object file, which differs from how function symbols appear when compiled using C.

To allow your C code to be used by C++ code, you must adapt the declarations to be read differently in C++ (or CUDA), in utils.h and VecAdd.h like so:

#ifdef __cplusplus
extern "C"
{
#endif

void PrintVec(float*, int);
// ... or whatever you're defining

#ifdef __cplusplus
}
#endif

Creating CUDA C library using CMake

1 Answers1