c++ - 在 Windows 上的 Qt Creator 中编译 Cuda 代码

Question

几天来，我一直在尝试让 Qt 项目文件在 32 位 Windows 7 系统上运行，我希望/需要在其中包含 Cuda 代码。事情的这种组合要么是如此简单以至于没有人愿意在网上发布一个例子，要么如此困难以至于没有人成功，似乎。不管怎样，我发现的唯一有用的论坛主题是 Linux 或 Mac 上的相同问题，或者 Windows 上的 Visual Studio。然而，所有这些都会产生各种不同的错误，无论是由于链接或冲突库，还是文件名中的空格或 Windows 版本的 Cuda SDK 中不存在的文件夹。有没有人.pro可以提供明确的文件来解决问题？

我的目标是用 Qt 风格的普通 C++ 代码和 Qt 4.8 库编译一个简单的程序，这些库引用了 .cu 文件中的几个 Cuda 模块。某种形式的东西：

TestCUDA \
    TestCUDA.pro
    main.cpp
    test.cu

score 15 · Accepted Answer

所以我终于设法组装了一个.pro适用于我的，可能适用于所有 Windows 系统的文件。以下是一个简单的测试程序，应该可以解决问题。以下是至少在我的系统上工作的小项目文件和测试程序。

文件系统如下所示：

TestCUDA \
    TestCUDA.pro
    main.cpp
    vectorAddition.cu

项目文件内容如下：

TARGET = TestCUDA

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

# Source files
SOURCES += src/main.cpp

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

# CUDA settings <-- may change depending on your system
CUDA_SOURCES += src/cuda/vectorAddition.cu
CUDA_SDK = "C:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.2/C"   # Path to cuda SDK install
CUDA_DIR = "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v4.2"            # 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_11           # 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

# The following library conflicts with something in Cuda
QMAKE_LFLAGS_RELEASE = /NODEFAULTLIB:msvcrt.lib
QMAKE_LFLAGS_DEBUG   = /NODEFAULTLIB:msvcrtd.lib

# 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
}

请注意QMAKE_LFLAGS_RELEASE = /NODEFAULTLIB:msvcrt.lib：我花了很长时间才弄清楚，但是这个库似乎与 Cuda 中的其他东西发生冲突，这会产生奇怪的链接警告和错误。如果有人对此有解释，并且可能有更漂亮的方法来解决这个问题，我想听听。

此外，由于 Windows 文件路径通常包含空格（NVIDIA 的 SDK 默认也包含空格），因此需要人为地在包含路径周围添加引号。同样，如果有人知道解决这个问题的更优雅的方法，我很想知道。

该main.cpp文件如下所示：

#include <cuda.h>
#include <builtin_types.h>
#include <drvapi_error_string.h>

#include <QtCore/QCoreApplication>
#include <QDebug>

// Forward declare the function in the .cu file
void vectorAddition(const float* a, const float* b, float* c, int n);

void printArray(const float* a, const unsigned int n) {
    QString s = "(";
    unsigned int ii;
    for (ii = 0; ii < n - 1; ++ii)
        s.append(QString::number(a[ii])).append(", ");
    s.append(QString::number(a[ii])).append(")");

    qDebug() << s;
}

int main(int argc, char* argv [])
{
    QCoreApplication(argc, argv);

    int deviceCount = 0;
    int cudaDevice = 0;
    char cudaDeviceName [100];

    unsigned int N = 50;
    float *a, *b, *c;

    cuInit(0);
    cuDeviceGetCount(&deviceCount);
    cuDeviceGet(&cudaDevice, 0);
    cuDeviceGetName(cudaDeviceName, 100, cudaDevice);
    qDebug() << "Number of devices: " << deviceCount;
    qDebug() << "Device name:" << cudaDeviceName;

    a = new float [N];    b = new float [N];    c = new float [N];
    for (unsigned int ii = 0; ii < N; ++ii) {
        a[ii] = qrand();
        b[ii] = qrand();
    }

    // This is the function call in which the kernel is called
    vectorAddition(a, b, c, N);

    qDebug() << "input a:"; printArray(a, N);
    qDebug() << "input b:"; printArray(b, N);
    qDebug() << "output c:"; printArray(c, N);

    if (a) delete a;
    if (b) delete b;
    if (c) delete c;
}

Cuda 文件vectorAddition.cu描述了一个简单的向量加法，如下所示：

#include <cuda.h>
#include <builtin_types.h>

extern "C"
__global__ void vectorAdditionCUDA(const float* a, const float* b, float* c, int n)
{
    int ii = blockDim.x * blockIdx.x + threadIdx.x;
    if (ii < n)
        c[ii] = a[ii] + b[ii];
}

void vectorAddition(const float* a, const float* b, float* c, int n) {
    float *a_cuda, *b_cuda, *c_cuda;
    unsigned int nBytes = sizeof(float) * n;
    int threadsPerBlock = 256;
    int blocksPerGrid   = (n + threadsPerBlock - 1) / threadsPerBlock;

    // allocate and copy memory into the device
    cudaMalloc((void **)& a_cuda, nBytes);
    cudaMalloc((void **)& b_cuda, nBytes);
    cudaMalloc((void **)& c_cuda, nBytes);
    cudaMemcpy(a_cuda, a, nBytes, cudaMemcpyHostToDevice);
    cudaMemcpy(b_cuda, b, nBytes, cudaMemcpyHostToDevice);

    vectorAdditionCUDA<<<blocksPerGrid, threadsPerBlock>>>(a_cuda, b_cuda, c_cuda, n);

    // load the answer back into the host
    cudaMemcpy(c, c_cuda, nBytes, cudaMemcpyDeviceToHost);

    cudaFree(a_cuda);
    cudaFree(b_cuda);
    cudaFree(c_cuda);
}

如果你让它工作，那么更复杂的例子是不言而喻的，我认为。

编辑（24-1-2013）：我添加了QMAKE_LFLAGS_DEBUG = /NODEFAULTLIB:msvcrtd.lib和CONFIG(debug)带有额外D_DEBUG标志的，这样它也可以在调试模式下编译。

score 1 · Accepted Answer

使用 msvc 2010 我发现链接器不接受 -l 参数，但是 nvcc 需要它。因此，我在 .pro 文件中做了一个简单的更改：

# Add the necessary libraries
CUDA_LIBS = cuda cudart
# The following makes sure all path names (which often include spaces) are put between quotation marks
CUDA_INC = $$join(INCLUDEPATH,'" -I"','-I"','"')
# LIBRARIES IN FORMAT NEEDED BY NVCC
NVCC_LIBS = $$join(CUDA_LIBS,' -l','-l', '')
# LIBRARIES IN FORMAT NEEDED BY VISUAL C++ LINKER
LIBS += $$join(CUDA_LIBS,'.lib ', '', '.lib')

和 nvcc 命令（发行版）：

cuda.commands = $$CUDA_DIR/bin/nvcc.exe $$NVCC_OPTIONS $$CUDA_INC $$NVCC_LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}

插入了 $$NVCC_LIBS 而不是 $$LIBS。整个 .pro 文件，对我有用：

QT       += core

QT       -= gui

TARGET = TestCUDA
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

# 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 4.2/C"   # Path to cuda SDK install
CUDA_DIR = "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v5.0"           # 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_11           # 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


# The following library conflicts with something in Cuda
QMAKE_LFLAGS_RELEASE = /NODEFAULTLIB:msvcrt.lib
QMAKE_LFLAGS_DEBUG   = /NODEFAULTLIB:msvcrtd.lib

# Add the necessary libraries
CUDA_LIBS = cuda cudart
# The following makes sure all path names (which often include spaces) are put between quotation marks
CUDA_INC = $$join(INCLUDEPATH,'" -I"','-I"','"')
NVCC_LIBS = $$join(CUDA_LIBS,' -l','-l', '')
LIBS += $$join(CUDA_LIBS,'.lib ', '', '.lib')

# 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 $$NVCC_LIBS --machine $$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 $$CUDA_INC $$NVCC_LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda
}

我还添加了一些必要的声明，即 QT += core 以使应用程序正常工作，并且还删除了 SDK 部分，在这种情况下我没有发现它有用。

score 0 · Accepted Answer

我试过这个组合来工作。由于我的项目中有许多依赖项，无法使其工作。我的最终解决方案是在 Windows 1 上将应用程序分成两个单独的应用程序）

在 VC 中开发并在 Windows 中作为服务/DLL 运行的 CUDA 应用程序
在 QT 中开发的 GUI 界面并使用 DLL 进行 CUDA 相关任务。

希望它可以节省其他人的一些时间

c++ - 在 Windows 上的 Qt Creator 中编译 Cuda 代码

3 回答 3

Related

Reference