OpenCV/CUDA项目CMake报错：缺失CUDAConfig.cmake

多次安装、卸载CUDA后，/usr/local/cuda/bin目录缺少CUDA 12.2应有的卸载脚本，且尝试多数论坛方法后，仍无法解决CMake配置问题，相关信息如下：

平台信息 #

• HP Omen 15
• Linux Ubuntu 22.04
• GeForce RTX 2070
• NVIDIA-SMI 535.54.03 - Driver Version: 535.54.03
• CUDA Version: 12.2

CMakeLists.txt内容 #

cmake_minimum_required(VERSION 3.27)
project(synchronisation_buffer)

# Set C++ Standards
set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)

# FIND LIBRARIES
find_package(OpenCV REQUIRED)
message(STATUS "OpenCV library status:")
message(STATUS "    version: ${OpenCV_VERSION}")
message(STATUS "    libraries: ${OpenCV_LIBS}")
message(STATUS "    include path: ${OpenCV_INCLUDE_DIRS}")

find_package(CUDA REQUIRED)
message(STATUS "    libraries: ${CUDA_LIBS}")

# INCLUDE HEADERS
include_directories(${OpenCV_INCLUDE_DIRS})
# Target the only executable
add_executable(${PROJECT_NAME}
    src/main.cpp 
    src/main.hpp 
    src/buffer.hpp 
)

.bashrc配置内容 #

# CUDA Bash links
# export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}$ 
# export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export PATH=$PATH:/usr/local/cuda/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib:/usr/local/lib
export CPLUS_INCLUDE_PATH=/usr/local/cuda/include

/usr/local目录内容 #

XXXX@XXXX-OMEN-Laptop-15-ek0xxx:/usr/local$ ls
bin cuda cuda-12 cuda-12.2 doc etc games include lib man sbin share src

/usr/local/cuda/bin目录内容 #

XXXX@XXXX-OMEN-Laptop-15-ek0xxx:/usr/local/cuda/bin$ ls
bin2c cudafe++ cuobjdump nsight_ee_plugins_manage.sh nsys-ui nvdisasm nvvp
computeprof cuda-gdb fatbinary nsight-sys nvcc nvlink ptxas
compute-sanitizer cuda-gdbserver ncu nsys __nvcc_device_query nvprof
crt cu++filt ncu-ui nsys-exporter nvcc.profile nvprune

报错信息 #

-- The C compiler identification is GNU 11.3.0
-- The CXX compiler identification is GNU 11.3.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
CMake Error at /usr/local/lib/cmake/opencv4/OpenCVConfig.cmake:86 (find_package):
  By not providing "FindCUDA.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "CUDA", but
  CMake did not find one.

  Could not find a package configuration file provided by "CUDA" (requested
  version 12.2) with any of the following names:

    CUDAConfig.cmake
    cuda-config.cmake

  Add the installation prefix of "CUDA" to CMAKE_PREFIX_PATH or set
  "CUDA_DIR" to a directory containing one of the above files.  If "CUDA"
  provides a separate development package or SDK, be sure it has been
  installed.
Call Stack (most recent call first):
  /usr/local/lib/cmake/opencv4/OpenCVConfig.cmake:108 (find_host_package)
  CMakeLists.txt:11 (find_package)


-- Configuring incomplete, errors occurred!
make: *** No targets specified and no makefile found.  Stop.

解决方案 #

1. 检查CUDA安装完整性 #

首先确认CUDA 12.2的CMake配置文件是否存在：

ls /usr/local/cuda-12.2/lib64/cmake/CUDA/CUDAConfig.cmake

如果文件不存在，说明CUDA Toolkit安装不完整，需要重新下载CUDA 12.2 runfile安装包，运行时勾选"CUDA Toolkit"选项（不要只安装驱动），确保完整安装。

2. 替换CMake中的CUDA查找方式 #

CMake 3.10+已弃用旧的find_package(CUDA)模块，推荐使用CUDAToolkit组件，修改CMakeLists.txt：

# 替换原有的find_package(CUDA REQUIRED)
find_package(CUDAToolkit REQUIRED)
message(STATUS "    CUDA libraries: ${CUDAToolkit_LIBRARIES}")
message(STATUS "    CUDA include path: ${CUDAToolkit_INCLUDE_DIRS}")

# 同时需要将CUDA库链接到目标
target_link_libraries(${PROJECT_NAME} PRIVATE ${CUDAToolkit_LIBRARIES})

3. 手动指定CUDA_DIR路径 #

如果确认CUDAConfig.cmake存在，可在CMake命令行中手动指定路径：

cmake .. -DCUDA_DIR=/usr/local/cuda-12.2/lib64/cmake/CUDA

或者在CMakeLists.txt开头添加：

set(CUDA_DIR "/usr/local/cuda-12.2/lib64/cmake/CUDA" CACHE PATH "Path to CUDAConfig.cmake directory")

4. 修复CUDA软链接 #

确保/usr/local/cuda软链接指向正确的CUDA版本：

# 先删除旧链接
sudo rm /usr/local/cuda
# 建立新链接指向cuda-12.2
sudo ln -s /usr/local/cuda-12.2 /usr/local/cuda

5. 补充环境变量 #

在.bashrc中添加CMAKE_PREFIX_PATH：

export CMAKE_PREFIX_PATH=/usr/local/cuda-12.2:$CMAKE_PREFIX_PATH

执行source ~/.bashrc生效后，再重新运行CMake。

内容的提问来源于stack exchange，提问作者CXX_NVCC