c++ - 编译cuda文件报错: "runtime library" mismatch value 'MDd_DynamicDebug' doesn't match value 'MTd_StaticDebug' in vectorAddition_cuda. o

标签 c++ qt cuda

我尝试在 Qt 5.2 和 MSVC2012 环境下编译一个 cuda 文件。 在开始我的项目之前，我仔细阅读了问题并回复:Compiling Cuda code in Qt Creator on Windows . 但是即使我简单地复制代码并在qt中生成2个文件，仍然会弹出一些错误

/main.cpp
/vectorAddition.cu

错误是:

error LNK2038: mismatch detected for 'RuntimeLibrary': value 'MDd_DynamicDebug' doesn't match value 'MTd_StaticDebug' in vectorAddition_cuda.o

我完全不知道如何修复这个错误。我在网上阅读了一些关于修复 MSVC2012 中的不匹配问题的文章。但很少有人提到如何在 Qt .pro 文件中做到这一点。有没有人碰巧遇到这个问题？

我这里也附上了pro文件(根据我自己的目录稍作修改)

#-------------------------------------------------
#
# Project created by QtCreator 2014-10-05T13:22:19
#
#-------------------------------------------------

QT       += core gui

greaterThan(QT_MAJOR_VERSION, 4): QT += widgets

TARGET = untitled
TEMPLATE = app


SOURCES += main.cpp

HEADERS  += mainwindow.h

FORMS    += mainwindow.ui


# Define output directories
DESTDIR = debug
OBJECTS_DIR = debug/obj
CUDA_OBJECTS_DIR = debug/obj

# Source files

# This makes the .cu files appear in your project
OTHER_FILES += vectorAddition.cu

# CUDA settings <-- may change depending on your system
CUDA_SOURCES += vectorAddition.cu
CUDA_SDK = "C:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.0/C"   # Path to cuda SDK install
CUDA_DIR = "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v6.5"            # Path to cuda toolkit install
SYSTEM_NAME = Win32         # Depending on your system either 'Win32', 'x64', or 'Win64'
SYSTEM_TYPE = 32            # '32' or '64', depending on your system
CUDA_ARCH = sm_30           # Type of CUDA architecture, for example 'compute_10', 'compute_11', 'sm_10'
NVCC_OPTIONS = --use_fast_math

# include paths
INCLUDEPATH += $$CUDA_DIR/include \
               $$CUDA_SDK/common/inc/ \
               $$CUDA_SDK/../shared/inc/

# library directories
QMAKE_LIBDIR += $$CUDA_DIR/lib/$$SYSTEM_NAME \
                $$CUDA_SDK/common/lib/$$SYSTEM_NAME \
                $$CUDA_SDK/../shared/lib/$$SYSTEM_NAME
# Add the necessary libraries
LIBS += -lcuda -lcudart
# MSVCRT link option (static or dynamic, it must be the same with your Qt SDK link option)
MSVCRT_LINK_FLAG_DEBUG = "/MDd"
MSVCRT_LINK_FLAG_RELEASE = "/MD"

# The following makes sure all path names (which often include spaces) are put between quotation marks
CUDA_INC = $$join(INCLUDEPATH,'" -I"','-I"','"')

# Configuration of the Cuda compiler
CONFIG(debug, debug|release) {
    # Debug mode
    cuda_d.input = CUDA_SOURCES
    cuda_d.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
    cuda_d.commands = $$CUDA_DIR/bin/nvcc.exe -D_DEBUG $$NVCC_OPTIONS $$CUDA_INC $$LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda_d.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda_d
}
else {
    # Release mode
    cuda.input = CUDA_SOURCES
    cuda.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
    cuda.commands = $$CUDA_DIR/bin/nvcc.exe $$NVCC_OPTIONS $$CUDA_INC $$LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda
}

这是我的日志

13:58:14: Running steps for project untitled...
13:58:14: Configuration unchanged, skipping qmake step.
13:58:14: Starting: "C:\Qt\Qt5.2.0\Tools\QtCreator\bin\jom.exe" 
    C:\Qt\Qt5.2.0\Tools\QtCreator\bin\jom.exe -f Makefile.Debug
    "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v6.5\bin\nvcc.exe" -D_DEBUG --use_fast_math -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v6.5/include" -I"C:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.0/C/common/inc/" -I"C:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.0/C/../shared/inc/" -lcuda -lcudart --machine 32 -arch=sm_30 -c -o debug\obj\vectorAddition_cuda.obj ..\untitled\vectorAddition.cu
vectorAddition.cu
    echo 1 /* CREATEPROCESS_MANIFEST_RESOURCE_ID */ 24 /* RT_MANIFEST */ "debug\\untitled.exe.embed.manifest">debug\untitled.exe_manifest.rc
    if not exist debug\untitled.exe if exist debug\untitled.exe.embed.manifest del debug\untitled.exe.embed.manifest
    if exist debug\untitled.exe.embed.manifest copy /Y debug\untitled.exe.embed.manifest debug\untitled.exe_manifest.bak
    link /NOLOGO /DYNAMICBASE /NXCOMPAT /NODEFAULTLIB:msvcrtd.lib /SUBSYSTEM:WINDOWS "/MANIFESTDEPENDENCY:type='win32' name='Microsoft.Windows.Common-Controls' version='6.0.0.0' publicKeyToken='6595b64144ccf1df' language='*' processorArchitecture='*'" /MANIFEST /MANIFESTFILE:debug\untitled.exe.embed.manifest /OUT:debug\untitled.exe @C:\Users\JYOU\AppData\Local\Temp\untitled.exe.4748.3744.jom
main.obj : error LNK2038: mismatch detected for 'RuntimeLibrary': value 'MDd_DynamicDebug' doesn't match value 'MTd_StaticDebug' in vectorAddition_cuda.obj
moc_mainwindow.obj : error LNK2038: mismatch detected for 'RuntimeLibrary': value 'MDd_DynamicDebug' doesn't match value 'MTd_StaticDebug' in vectorAddition_cuda.obj
LINK : fatal error LNK1181: cannot open input file 'C:\Program.obj'
jom: C:\Qt\Qt5.2.0\Tools\QtCreator\bin\build-untitled-Desktop_Qt_5_2_0_MSVC2012_32bit-Debug\Makefile.Debug [debug\untitled.exe] Error 1181
jom: C:\Qt\Qt5.2.0\Tools\QtCreator\bin\build-untitled-Desktop_Qt_5_2_0_MSVC2012_32bit-Debug\Makefile [debug] Error 2
13:58:18: The process "C:\Qt\Qt5.2.0\Tools\QtCreator\bin\jom.exe" exited with code 2.
Error while building/deploying project untitled (kit: Desktop Qt 5.2.0 MSVC2012 32bit)
When executing step 'Make'
13:58:18: Elapsed time: 00:04.

“问题重复”的解释:

我也看了文章Mismatch Detected for 'RuntimeLibrary' 但不同的是，我的问题是关于在 Qt 中设置 NVCC，它比在 MSVC 环境中更复杂。即使/MDd 由以下代码设置，如果没有正确设置 NVCC，cu 文件仍然无法编译。运行时库的简单更改并不能完全解决我的问题。正是NVCC的设定最终成就了它。 

# MSVCRT link option (static or dynamic, it must be the same with your Qt SDK link option)
MSVCRT_LINK_FLAG_DEBUG = "/MDd"
MSVCRT_LINK_FLAG_RELEASE = "/MD"

最佳答案

我通过阅读另一篇文章解决了我的问题 http://robot9.me/cuda_qtcreator_windows/

作为问题Mismatch Detected for 'RuntimeLibrary'建议。 MSVCRT 链接选项需要保持一致。所以在 NVCC 配置之前添加了 /MDd 设置。

但关键是要正确设置cuda.commands。不同于问题:Compiling Cuda code in Qt Creator on Windows (我试过那个配置但对我不起作用) 我的 nvcc 是这样设置的:

   # MSVCRT link option (static or dynamic, it must be the same with your Qt SDK link option)
    MSVCRT_LINK_FLAG_DEBUG = "/MDd"
    MSVCRT_LINK_FLAG_RELEASE = "/MD"

   CONFIG(debug, debug|release) {
    #Debug settings
    # Debug mode
    cuda_d.input    = CUDA_SOURCES
    cuda_d.output   = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.obj
    cuda_d.commands = $$CUDA_DIR/bin/nvcc.exe -D_DEBUG $$NVCC_OPTIONS $$CUDA_INC $$LIBS \
                      --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH \
                      --compile -cudart static -g -DWIN32 -D_MBCS \
                      -Xcompiler "/wd4819,/EHsc,/W3,/nologo,/Od,/Zi,/RTC1" \
                      -Xcompiler $$MSVCRT_LINK_FLAG_DEBUG \
                      -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda_d.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda_d
}
else {
     # Release settings
     cuda.input    = CUDA_SOURCES
     cuda.output   = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.obj
     cuda.commands = $$CUDA_DIR/bin/nvcc.exe $$NVCC_OPTIONS $$CUDA_INC $$LIBS \
                    --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH \
                    --compile -cudart static -DWIN32 -D_MBCS \
                    -Xcompiler "/wd4819,/EHsc,/W3,/nologo,/O2,/Zi" \
                    -Xcompiler $$MSVCRT_LINK_FLAG_RELEASE \
                    -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
     cuda.dependency_type = TYPE_C
     QMAKE_EXTRA_COMPILERS += cuda
}

虽然我还没有完全理解一些cuda.commands，但它确实有效。 请引用http://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#axzz3FOBrgXpc有关 NVCC 配置的更多详细信息。

这是我的 .pro 文件的完整版本

#-------------------------------------------------
#
# Project created by QtCreator 2013-09-01T16:29:35
#
#-------------------------------------------------

QT       += core gui

greaterThan(QT_MAJOR_VERSION, 4): QT += widgets

TARGET = FFTW //your project name
TEMPLATE = app

HEADERS  += dialog.h \
    include/fftw3.h \   //header i need for my project 


SOURCES += main.cpp\
           dialog.cpp


INCLUDEPATH += $$PWD/
DEPENDPATH += $$PWD/

DESTDIR = debug           
OBJECTS_DIR = debug/obj           # directory where .obj files will be saved
CUDA_OBJECTS_DIR = debug/obj      # directory where .obj  of cuda file will be saved
# This makes the .cu files appear in your project

OTHER_FILES += intmapping.cu      # this is my cu file need to compile

# CUDA settings <-- may change depending on your system

CUDA_SOURCES += intmapping.cu     # let NVCC know which file you want to compile CUDA NVCC

CUDA_SDK = "C:\ProgramData\NVIDIA Corporation\NVIDIA GPU Computing SDK 4.0\C"   # Path to cuda SDK install
CUDA_DIR = "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v6.5"
SYSTEM_NAME = Win32         # Depending on your system either 'Win32', 'x64', or 'Win64'
SYSTEM_TYPE = 32            # '32' or '64', depending on your system
CUDA_ARCH = sm_20           # Type of CUDA architecture, for example 'compute_10', 'compute_11', 'sm_10'
NVCC_OPTIONS += --use_fast_math # default setting

# include paths

INCLUDEPATH += $$CUDA_DIR/include\
                $$CUDA_SDK/common/inc\
                $$CUDA_SDK/../shared/inc\


# library directories
QMAKE_LIBDIR += $$join(CUDA_DIR,'" -I"','-I"','"')/lib/$$SYSTEM_NAME\
                $$join(CUDA_SDK,'" -I"','-I"','"')/common/lib/$$SYSTEM_NAME\
                $$join(CUDA_SDK,'" -I"','-I"','"')/../shared/lib/$$SYSTEM_NAME

# Add the necessary libraries
CUDA_LIBS= -lcuda -lcudart -lcufft
#add quotation for those directories contain space (Windows required)
CUDA_INC +=$$join(INCLUDEPATH,'" -I"','-I"','"')

LIBS += $$CUDA_LIBS
#nvcc config
# MSVCRT link option (static or dynamic, it must be the same with your Qt SDK link option)
MSVCRT_LINK_FLAG_DEBUG = "/MDd"
MSVCRT_LINK_FLAG_RELEASE = "/MD"

CONFIG(debug, debug|release) {
    #Debug settings
    # Debug mode
    cuda_d.input    = CUDA_SOURCES
    cuda_d.output   = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.obj
    cuda_d.commands = $$CUDA_DIR/bin/nvcc.exe -D_DEBUG $$NVCC_OPTIONS $$CUDA_INC $$LIBS \
                      --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH \
                      --compile -cudart static -g -DWIN32 -D_MBCS \
                      -Xcompiler "/wd4819,/EHsc,/W3,/nologo,/Od,/Zi,/RTC1" \
                      -Xcompiler $$MSVCRT_LINK_FLAG_DEBUG \
                      -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda_d.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda_d
}
else {
     # Release settings
     cuda.input    = CUDA_SOURCES
     cuda.output   = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.obj
     cuda.commands = $$CUDA_DIR/bin/nvcc.exe $$NVCC_OPTIONS $$CUDA_INC $$LIBS \
                    --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH \
                    --compile -cudart static -DWIN32 -D_MBCS \
                    -Xcompiler "/wd4819,/EHsc,/W3,/nologo,/O2,/Zi" \
                    -Xcompiler $$MSVCRT_LINK_FLAG_RELEASE \
                    -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
     cuda.dependency_type = TYPE_C
     QMAKE_EXTRA_COMPILERS += cuda
}

FORMS += \
     dialog.ui

win32: LIBS += -L$$PWD/lib/ -llibfftw3f-3 #library i need in my project

INCLUDEPATH += $$PWD/lib
DEPENDPATH += $$PWD/lib

关于c++ - 编译cuda文件报错: "runtime library" mismatch value 'MDd_DynamicDebug' doesn't match value 'MTd_StaticDebug' in vectorAddition_cuda. o，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/26205608/

上一篇：c++ - 如何测量共享内存中 boost 进程间 vector 的大小？

下一篇：c++ - QPainter.drawText() SIGSEGV 段错误

相关文章：

qt - X-Windows、wxWidgets、GTK/Qt 和 OpenGL 有什么区别？

cuda - 如何在Windows上模拟CUDA

c - 我是否正确使用常量内存？

cmake - 为什么需要分离编译？

c++ - QTimer，参数化槽

c++ - 多模态图像对齐问题

c++ - 使用特定的父级从 boost 图制作子图

c++ - "not, and, or, not_eq.."是 C++ 标准的一部分吗？ (以及为什么可以在代码中使用或避免使用它们？)

Qt 可编辑 QGraphicsTextItem 验证文本并在更改时发出信号

c++ - 引用为类成员初始化

©2024 IT工具网  联系我们