ncnn - ubuntu调用vulkan加速ncnn是的CMakeLists.txt和bashrc配置

概述

ncnn使用vulkan来加速时,需要先安装vulkan,然后再使用vulkan来编译ncnn,最后才是使用编译好的ncnn来推理模型。。

这里的vulkan再Ubuntu中通常是直接解压安装,然后配置环境变量如下:

export VULKAN_SDK=/home/fwq/softs/1.2.182.0/x86_64

1.2.182.0 是我使用的vulkan版本。。

配置完上述路径后,source ~/.bashrc。

这时,编译ncnn 时 依赖 vulkan就成功通过了,但使用CMakeLists.txt 调用ncnn时报错:

CMake Error at /usr/local/share/cmake-3.23/Modules/FindPackageHandleStandardArgs.cmake:230 (message):
 Could NOT find Vulkan (missing: Vulkan_LIBRARY Vulkan_INCLUDE_DIR) (found
 version "")

Call Stack (most recent call first):
 /usr/local/share/cmake-3.23/Modules/FindPackageHandleStandardArgs.cmake:594 (_FPHSA_FAILURE_MESSAGE)
 /usr/local/share/cmake-3.23/Modules/FindVulkan.cmake:159 (find_package_handle_standard_args)
 /home/fwq/ncnn/lib/cmake/ncnn/ncnnConfig.cmake:18 (find_package)

解决1(失败)

添加环境变量配置:

##vulkan
export VULKAN_SDK=/home/fwq/softs/1.2.182.0/x86_64

export Vulkan_LIBRARY=/home/fwq/softs/1.2.182.0/x86_64/lib

export Vulkan_INCLUDE_DIR=/home/fwq/softs/1.2.182.0/x86_64/include

export vulkan_BIN=/home/fwq/softs/1.2.182.0/x86_64/bin

export PATH=$PATH:$Vulkan_LIBRARY:$Vulkan_INCLUDE_DIR:$vulkan_BIN

然后重新导入CMakeLIsts.txt,但还是失败

解决2(成功)

这个解决1中的并没有注释-- 后续可以测试一下,注释过后有没有影响

在CMakeLists.txt中添加:

set(Vulkan_LIBRARY /home/fwq/softs/1.2.182.0/x86_64/lib)

set(Vulkan_INCLUDE_DIR /home/fwq/softs/1.2.182.0/x86_64/include)

完整导入ncnn的CMakeLists.txt示例

set(Vulkan_LIBRARY /home/fwq/softs/1.2.182.0/x86_64/lib)
set(Vulkan_INCLUDE_DIR /home/fwq/softs/1.2.182.0/x86_64/include)set(ncnn_DIR "/home/fwq/ncnn/lib/cmake/ncnn")


set(ncnn_INCLUDE_DIRS /home/fwq/ncnn/include)
find_package(ncnn REQUIRED)

# 添加openmp的编译选项
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fopenmp")

link_directories(
       ${ncnn_INCLUDE_DIRS}
)


add_library(yolov5lite_ncnn SHARED  ${yolov5lite_ncnn_cpp})
target_link_libraries(yolov5lite_ncnn
        ${OpenCV_LIBS}
        ncnn
  )

target_include_directories(yolov5lite_ncnn PUBLIC
        ${CMAKE_SOURCE_DIR}/3rdpart
        ${CMAKE_SOURCE_DIR}/3rdpart/ncnn_include

        ${CMAKE_CURRENT_LIST_DIR}/Export
        ${CMAKE_CURRENT_LIST_DIR}/Include

        )张无门槛流量券(+1000曝光),0点自动发

你可能感兴趣的:(模型部署-ncnn,ubuntu,linux,运维)