因工作需要,我要在ubuntu20.04上编译cuda,且用到Qt。编译过程的要点如下:
1 扩展名
cuda的头文件不需要是cuh为扩展名,普通的.h文件即可
2 编译配置
在pro文件里,你会看到这两句:
SYSTEM_TYPE = 64
CUDA_ARCH = compute_61
64表示64位程序;compute_61指的是nvidia显卡的compute capability,与显卡型号有关。
比如Tesla P40 和Tesla P4都对应6.1,那么CUDA_ARCH就是compute_61
此外,在安装了cuda驱动后你会在实例程序里面找到1_Utilities/deviceQuery。运行该程序,也可以看到其计算能力。详见[NVIDIA] Ububtu 获取 CUDA_ARCH_hanjialeOK的博客-CSDN博客
3 在pro文件里把相关库的路径写上
QT -= gui #+= core gui
CONFIG += c++11
CONFIG -= app_bundle
# You can make your code fail to compile if it uses deprecated APIs.
# In order to do so, uncomment the following line.
#DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0x060000 # disables all the APIs deprecated before Qt 6.0.0
SOURCES += \
main.cpp
HEADERS += \
bilinear.h
# Default rules for deployment.
qnx: target.path = /tmp/$${TARGET}/bin
else: unix:!android: target.path = /opt/$${TARGET}/bin
!isEmpty(target.path): INSTALLS += target
LIBS += \
-L"/usr/local/lib" \
-L"/usr/local/cuda/lib64" \
-lcudart -lcufft \
-L"/usr/lib/gcc/x86_64-linux-gnu/7" \
-lgomp
CUDA_SOURCES += bilinear.cu
CUDA_SDK = "/usr/local/cuda/"
CUDA_DIR = "/usr/local/cuda/"
SYSTEM_NAME = linux
SYSTEM_TYPE = 64
CUDA_ARCH = compute_61
NVCC_OPTIONS = --use_fast_math
INCLUDEPATH += $$PWD/. \
$$CUDA_DIR/include
QMAKE_LIBDIR += $$CUDA_DIR/lib64/
CUDA_OBJECTS_DIR = ./
CUDA_LIBS = cudart cufft
CUDA_INC = $$join(INCLUDEPATH, '" -I"', '-I"', '"')
NVCC_LIBS = $$join(CUDA_LIBS, ' -l', '-l', '')
DEPENDPATH += .
CONFIG(debug, debug|release) {
cuda_d.input = CUDA_SOURCES
cuda_d.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
cuda_d.commands = $$CUDA_DIR/bin/nvcc -D_DEBUG $$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 {
cuda.input = CUDA_SOURCES
cuda.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
cuda.commands = $$CUDA_DIR/bin/nvcc $$NVCC_OPTIONS $$CUDA_INC $$NVCC_LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -O3 -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
cuda.dependency_type = TYPE_C
QMAKE_EXTRA_COMPILERS += cuda
}
DISTFILES += \
bilinear.cu
由于程序里还用到了omp,所以pro文件里也加入了-lgomp(后来omp部分的代码删掉了,但是pro文件的依赖还留着)
具体代码见github
采用Qt编译cuda的实例代