MAC에서 QT + CUDA 사용하기

1. 준비사항

∙ CUDA 사용이 가능한 GPU (CUDA 사용 가능 하드웨어 목록)

∙ Mac OS X 10.8 or later 

∙ gcc 또는 Clang 컴파일러와 툴체인이 Xcode에 설치되어 있을 것

∙ Command Line Tools 패키지 설치 필요

∙ the NVIDIA CUDA Toolkit 설치할 것(CUDA Download page)

∙ 환경변수 설정 (매우 중요)

export PATH=/Developer/NVIDIA/CUDA-6.0/bin:$PATH 

export DYLD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-6.0/lib:$DYLD_LIBRARY_PATH

2. 설치 확인

∙ 드라이버 설치 확인

kextstat | grep -i cuda

∙ 컴파일러 설치 확인

nvcc -V

∙ 예제프로그램 컴파일 후 실행

make -C 1_Utilities/deviceQuery
cd bin/x86_64/darwin/release
./deviceQuery

3. *QT Creator에서 설정하기

∙ .pro (맨 아래 네 줄이 가장 중요한 포인트임)

QT       += core

QT       -= gui

TARGET = cudatest    # depending on the project

CONFIG   += console

CONFIG   -= app_bundle

TEMPLATE = app

# Basic .pro configuration

SOURCES += \

    main.cpp

# This makes the .cu files appear in your project

OTHER_FILES += main.cu

# Cuda sources

CUDA_SOURCES += main.cu

# CUDA settings (depending on the system)

CUDA_DIR = /usr/local/cuda      # Path to cuda toolkit install

# nvcc flags (ptxas option verbose is always useful)

NVCCFLAGS = --compiler-options -fno-strict-aliasing -use_fast_math --ptxas-options=-v

# include paths

INCLUDEPATH += $$CUDA_DIR/include

# lib dirs

QMAKE_LIBDIR += $$CUDA_DIR/lib

# libs

LIBS += -lcudart

# join the includes in a line

CUDA_INC = $$join(INCLUDEPATH,' -I','-I',' ')

# Prepare the extra compiler configuration

cuda.input = CUDA_SOURCES

cuda.output = ${OBJECTS_DIR}${QMAKE_FILE_BASE}_cuda.o

cuda.commands = $$CUDA_DIR/bin/nvcc -gencode arch=compute_10,code=sm_10 -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_32,code=sm_32 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_50,code=compute_50 -c $$NVCCFLAGS $$CUDA_INC $$LIBS ${QMAKE_FILE_NAME} -o ${QMAKE_FILE_OUT}

cuda.dependency_type = TYPE_C

cuda.depend_command = $$CUDA_DIR/bin/nvcc -gencode arch=compute_10,code=sm_10 -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_32,code=sm_32 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_50,code=compute_50 $$CUDA_INC $$NVCCFLAGS ${QMAKE_FILE_NAME}

# Tell Qt that we want add more stuff to the Makefile

QMAKE_EXTRA_COMPILERS += cuda

QMAKE_MACOSX_DEPLOYMENT_TARGET = 10.9

QMAKE_CXXFLAGS += -stdlib=libstdc++

QMAKE_LFLAGS += -stdlib=libstdc++ -rpath $$CUDA_DIR/lib

∙ main.cpp

#include <QtCore/QCoreApplication>

#include <cuda.h>

using namespace std;

void run();

int main(int argc, char *argv[])

{

    QCoreApplication a(argc, argv);

    run();

    return a.exec();

}

∙ main.cu

#include <iostream>

#include <string>

#include "cuda_runtime.h"

#include "device_launch_parameters.h"

#define SIZE 1024

using namespace std;

__global__ void VectorAdd(int *a, int *b, int *c, int n)

{

    int i = threadIdx.x;

    if( i < n )

        c[i] = a[i] + b[i];

}

void run()

{

    int *a, *b, *c;

    int *d_a, *d_b, *d_c;

    a = (int *)malloc( SIZE * sizeof(int) );

    b = (int *)malloc( SIZE * sizeof(int) );

    c = (int *)malloc( SIZE * sizeof(int) );

    cudaMalloc( &d_a, SIZE * sizeof(int) );

    cudaMalloc( &d_b, SIZE * sizeof(int) );

    cudaMalloc( &d_c, SIZE * sizeof(int) );

    int i;

    for( i=0; i<SIZE; ++i )

    {

        a[i] = i;

        b[i] = i;

        c[i] = 0;

    }

    cudaMemcpy( d_a, a, SIZE * sizeof(int), cudaMemcpyHostToDevice );

    cudaMemcpy( d_b, b, SIZE * sizeof(int), cudaMemcpyHostToDevice );

    cudaMemcpy( d_c, c, SIZE * sizeof(int), cudaMemcpyHostToDevice );

//    VectorAdd( a, b, c, SIZE );

    VectorAdd<<< 1, SIZE >>>(d_a, d_b, d_c, SIZE);

    cudaMemcpy( a, d_a, SIZE * sizeof(int), cudaMemcpyDeviceToHost );

    cudaMemcpy( b, d_b, SIZE * sizeof(int), cudaMemcpyDeviceToHost );

    cudaMemcpy( c, d_c, SIZE * sizeof(int), cudaMemcpyDeviceToHost );

    for( i=0; i<10; ++i )

        printf( "c[%d] = %d\n", i, c[i] );

    free(a);

    free(b);

    free(c);

    cudaFree(d_a);

    cudaFree(d_b);

    cudaFree(d_c);

}

4. 테스트 결과

SPECIAL THANKS TO 문상환

출처 : http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-mac-os-x