让dlib库也飞起来 | Ubuntu18.04下联合编译Dlib19.21和CUDA10.1

一、一些需要的东西：

• Ubuntu 18.04
• CUDA 10.1
• cuDNN 7.6.5
• g++ 7.5 (并不需要降级到5.0)
• cmake 3.18.2
• python 3.6.5
• Anaconda 3-5.2

 二、下载项目文件

git clone https://github.com/davisking/dlib.git
cd dlib
mkdir build
cd build

三、编译

DLIB_USE_CUDA=1 支持CUDA

                              =0 不支持CUDA

USE_AVX_INSTRUCTIONS AVX指令集的CPU加速

cmake .. -DDLIB_USE_CUDA=1 -DUSE_AVX_INSTRUCTIONS=1

会显示一些日志：

-- The C compiler identification is GNU 7.5.0
-- The CXX compiler identification is GNU 7.5.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Using CMake version: 3.18.2
-- Compiling dlib version: 19.21.99
-- Enabling AVX instructions
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found X11: /usr/include   
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so - found
-- Looking for gethostbyname
-- Looking for gethostbyname - found
-- Looking for connect
-- Looking for connect - found
-- Looking for remove
-- Looking for remove - found
-- Looking for shmat
-- Looking for shmat - found
-- Looking for IceConnectionNumber in ICE
-- Looking for IceConnectionNumber in ICE - found
-- Found system copy of libpng: /usr/lib/x86_64-linux-gnu/libpng.so;/usr/lib/x86_64-linux-gnu/libz.so
-- Found system copy of libjpeg: /usr/lib/x86_64-linux-gnu/libjpeg.so
-- Searching for BLAS and LAPACK
-- Searching for BLAS and LAPACK
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1") 
-- Checking for module 'cblas'
--   No package 'cblas' found
-- Checking for module 'lapack'
--   No package 'lapack' found
-- Looking for cblas_ddot
-- Looking for cblas_ddot - not found
-- Looking for sys/types.h
-- Looking for sys/types.h - found
-- Looking for stdint.h
-- Looking for stdint.h - found
-- Looking for stddef.h
-- Looking for stddef.h - found
-- Check size of void*
-- Check size of void* - done
 *****************************************************************************
 *** No BLAS library found so using dlib's built in BLAS.  However, if you ***
 *** install an optimized BLAS such as OpenBLAS or the Intel MKL your code ***
 *** will run faster.  On Ubuntu you can install OpenBLAS by executing:    ***
 ***    sudo apt-get install libopenblas-dev liblapack-dev                 ***
 *** Or you can easily install OpenBLAS from source by downloading the     ***
 *** source tar file from http://www.openblas.net, extracting it, and      ***
 *** running:                                                              ***
 ***    make; sudo make install                                            ***
 *****************************************************************************
-- Found CUDA: /usr/local/cuda (found suitable version "10.1", minimum required is "7.5") 
-- Looking for cuDNN install...
-- Found cuDNN: /usr/local/cuda/lib64/libcudnn.so
-- Building a CUDA test project to see if your compiler is compatible with CUDA...
-- Building a cuDNN test project to check if you have the right version of cuDNN installed...
-- Enabling CUDA support for dlib.  DLIB WILL USE CUDA
-- C++11 activated.
-- Configuring done
-- Generating done
-- Build files have been written to: /home/ubuntu/dlib/build

然后是编译：

cmake --build .

一些日志：

[  1%] Building NVCC (Device) object dlib/CMakeFiles/dlib.dir/cuda/dlib_generated_cusolver_dlibapi.cu.o
[  3%] Building NVCC (Device) object dlib/CMakeFiles/dlib.dir/cuda/dlib_generated_cuda_dlib.cu.o
Scanning dependencies of target dlib
[  4%] Building CXX object dlib/CMakeFiles/dlib.dir/base64/base64_kernel_1.cpp.o
[  6%] Building CXX object dlib/CMakeFiles/dlib.dir/bigint/bigint_kernel_1.cpp.o
[  7%] Building CXX object dlib/CMakeFiles/dlib.dir/bigint/bigint_kernel_2.cpp.o
[  9%] Building CXX object dlib/CMakeFiles/dlib.dir/bit_stream/bit_stream_kernel_1.cpp.o
[ 10%] Building CXX object dlib/CMakeFiles/dlib.dir/entropy_decoder/entropy_decoder_kernel_1.cpp.o
[ 12%] Building CXX object dlib/CMakeFiles/dlib.dir/entropy_decoder/entropy_decoder_kernel_2.cpp.o
[ 13%] Building CXX object dlib/CMakeFiles/dlib.dir/entropy_encoder/entropy_encoder_kernel_1.cpp.o
[ 15%] Building CXX object dlib/CMakeFiles/dlib.dir/entropy_encoder/entropy_encoder_kernel_2.cpp.o
[ 16%] Building CXX object dlib/CMakeFiles/dlib.dir/md5/md5_kernel_1.cpp.o
[ 18%] Building CXX object dlib/CMakeFiles/dlib.dir/tokenizer/tokenizer_kernel_1.cpp.o
[ 19%] Building CXX object dlib/CMakeFiles/dlib.dir/unicode/unicode.cpp.o
[ 21%] Building CXX object dlib/CMakeFiles/dlib.dir/test_for_odr_violations.cpp.o
[ 22%] Building CXX object dlib/CMakeFiles/dlib.dir/sockets/sockets_kernel_1.cpp.o
[ 24%] Building CXX object dlib/CMakeFiles/dlib.dir/bsp/bsp.cpp.o
[ 25%] Building CXX object dlib/CMakeFiles/dlib.dir/dir_nav/dir_nav_kernel_1.cpp.o
[ 27%] Building CXX object dlib/CMakeFiles/dlib.dir/dir_nav/dir_nav_kernel_2.cpp.o
[ 28%] Building CXX object dlib/CMakeFiles/dlib.dir/dir_nav/dir_nav_ex