在 QtCreator (VS 2010) 中设置 CUDA v7.0

Question

有人在 QtCreator 中成功使用了最新版本的 CUDA（即设置 .pro 文件）吗？如果是这样，你能和我分享一些简单的例子或教程吗？谢谢！

遇到的问题和尝试：

我在 QtCreator 中设置和编译一个简单的 CUDA 程序时遇到了麻烦。我已经阅读了一些旧教程（即CUDA with QT in Linux），但到目前为止我还没有运气。

在上面的教程示例中，我无法让 QtCreator 自行正确构建 .cu 文件，但我最终能够使用以下 nvcc 命令行编译 .cu 文件：

nvcc -m32 --gpu-architecture=sm_20 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin" -I"C:/PROGRA~1/NVIDIA GPU Computing Toolkit/CUDA/v7.0/include"  -lcudart -lcuda ..\CUDA_TEST_EXAMPLE\hello.cu -o ..\Release\release\obj\cuda_code_cuda.obj

于是生成了 cuda_code_cuda.obj。当我尝试在 Qt 中构建项目时，我收到一条错误消息：

error: LNK1181: cannot open input file 'C:\PROGRA~1\NVIDIA.obj'

我肯定没有正确链接到已编译的 cuda_code_cuda.obj 文件。我认为它与 CUDA 文件夹路径中的“空格”有关（双引号似乎没有多大帮助）

附加信息：我的 pro 文件如下所示：

QT       += core
QT       -= gui
TARGET = CUDA_TEST_EXAMPLE
CONFIG   += console
CONFIG   -= app_bundle
TEMPLATE = app

# Define output directories
DESTDIR = release
OBJECTS_DIR = release/obj/
CUDA_OBJECTS_DIR = release/cuda

# Source files
SOURCES += main.cpp
# Cuda sources
OTHER_FILES +=  cuda_code.cu

# CUDA settings <-- may change depending on your system
CUDA_SOURCES += cuda_code.cu
#CUDA_DIR = "C:/PROGRA~1/NVIDIA\ GPU\ Computing\ Toolkit/CUDA/v7.0"                    
# Path to header and libs files
INCLUDEPATH  += "C:\PROGRA~1\NVIDIA GPU Computing Toolkit\CUDA\v7.0\include"
QMAKE_LIBDIR += "C:\PROGRA~1\NVIDIA GPU Computing Toolkit\CUDA\v7.0\lib\Win32"    # Note I'm using a 64 bits Operating system
# libs used in your code
LIBS += -lcudart -lcuda

cuda.output = "C:\Users\johnyang\Documents\Release\release\obj\cuda_code_cuda.obj"

LIBS += "C:\Users\johnyang\Documents\Release\release\obj\cuda_code_cuda.obj"

## Tell Qt that we want add more stuff to the Makefile
QMAKE_EXTRA_COMPILERS += cuda

Answer 1

通过一些测试，我设法将 pro 文件设置为现在可以正确编译。我发布这个作为答案只是为了分享我所得到的。但是，我仍然很好奇这是否是最新版本的 CUDA 中的方法。如果有人能跟我确认一下就好了。

QT       += core
QT       -= gui

TARGET = CUDA_TEST_EXAMPLE
CONFIG   += console
CONFIG   -= app_bundle
TEMPLATE = app

QMAKE_LFLAGS_RELEASE =/NODEFAULTLIB:msvcrt.lib
QMAKE_LFLAGS_DEBUG =/NODEFAULTLIB:msvcrtd.lib

# Define output directories
DESTDIR = release
OBJECTS_DIR = release\obj
CUDA_OBJECTS_DIR = release/cuda

# Source files
SOURCES += main.cpp
# Cuda sources
#OTHER_FILES +=  cuda_code.cu

# CUDA settings <-- may change depending on your system
CUDA_SOURCES += cuda_code.cu
CUDA_DIR = "C:\PROGRA~1\NVIDIA~2\CUDA\v7.0"            # Path to cuda toolkit install

# Path to header and libs files
INCLUDEPATH  += $$CUDA_DIR/include
QMAKE_LIBDIR += $$CUDA_DIR/lib/Win32    # Note I'm using a 64 bits Operating system
# libs used in your code
LIBS += -lcudart -lcuda

## GPU architecture
CUDA_ARCH     = sm_20                # Yeah! I've a new device. Adjust with your compute capability
## Here are some NVCC flags I've always used by default.
#NVCCFLAGS     = --compiler-options -fno-strict-aliasing -use_fast_math --ptxas-options=-v

## Prepare the extra compiler configuration (taken from the nvidia forum - i'm not an expert in this part)
#CUDA_INC = -I"C:/PROGRA~1/NVIDIA~2/CUDA/v7.0/include"
CUDA_INC = $$join(INCLUDEPATH,' -I','-I',' ')

cuda.commands = $${CUDA_DIR}/bin/nvcc -m32 --gpu-architecture=$$CUDA_ARCH -c \
                $$CUDA_INC $$LIBS  ${QMAKE_FILE_NAME} -o ${QMAKE_FILE_OUT}
# nvcc error printout format ever so slightly different from gcc
# http://forums.nvidia.com/index.php?showtopic=171651

cuda.dependency_type = TYPE_C # there was a typo here. Thanks workmate!
cuda.depend_command = $$(CUDA_DIR)/bin/nvcc $$CUDA_INC ${QMAKE_FILE_NAME}

cuda.input = CUDA_SOURCES
cuda.output = ..\Release/$${OBJECTS_DIR}/${QMAKE_FILE_BASE}_cuda.obj

## Tell Qt that we want add more stuff to the Makefile
QMAKE_EXTRA_COMPILERS += cuda