淡泊明志，宁静致远

0%

显卡驱动及GPU环境配置

发表于 更新于 分类于 阅读次数： 评论数：
本文字数： 2.6k 阅读时长 ≈ 2 分钟

前言

caffe安装时，对GPU环境的搭建和配置。

本文环境为：ubuntu16.04，cuda8.0，cudnn5,opencv3.1。

安装驱动

  • ctrl+alt+f4，进入tty界面，关闭图形
1
$sudo service lightdm stop
  • 添加驱动源：
1
2
3
4
5
6
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt-get update
sudo apt-get install nvidia-367
sudo apt-get install nvidia-settings
sudo apt-get install nvidia-prime
sudo reboot
  • 添加显卡切换器(可选)：
1
2
3
sudo add-apt-repository ppa:nilarimogard/webupd8
sudo apt-get update
sudo apt-get install prime-indicator
1
nvidia-smi

cuda8.0安装

1
2
sudo service lightdm stop
sudo sh cuda_8.0.44_linux.run
1
sudo gedit ~/.bashrc
1
2
export PATH=/usr/local/cuda-8.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64:$LD_LIBRARY_PATH
1
source .bashrc
1
sudo gedit /etc/profile

添加如下内容

1
export PATH=/usr/local/cuda/bin:$PATH

创建新文档

1
sudo gedit /etc/ld.so.conf.d/cuda.conf

填写

1
/usr/local/cuda/lib64

使命令生效

1
sudo ldconfig
1
2
3
4
5
cd /usr/local/cuda-8.0/samples/1_Utilities/deviceQuery

sudo make

sudo ./deviceQuery

查看显卡信息

cudnn安装

1
2
cd ~/cuda/include/
sudo cp cudnn.h /usr/local/cuda/include/
1
2
cd ~/cuda/lib64/
sudo cp lib* /usr/local/cuda/lib64/
1
cd /usr/local/cuda/lib64/
1
2
3
sudo rm -rf libcudnn.so libcudnn.so.5 
sudo ln -s libcudnn.so.5.0.5 libcudnn.so.5
sudo ln -s libcudnn.so.5 libcudnn.so
1
nvcc -V

opencv-3.1安装

1
sudo apt-get install --assume-yes libopencv-dev build-essential cmake git libgtk2.0-dev pkg-config python-dev python-numpy libdc1394-22 libdc1394-22-dev libjpeg-dev libpng12-dev libtiff5-dev libjasper-dev libavcodec-dev libavformat-dev libswscale-dev libxine2-dev libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev libv4l-dev libtbb-dev libqt4-dev libfaac-dev libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev x264 v4l-utils unzip
1
sudo apt-get install build-essential cmake git
1
sudo apt-get install ffmpeg libopencv-dev libgtk-3-dev python-numpy python3-numpy libdc1394-22 libdc1394-22-dev libjpeg-dev libpng12-dev libtiff5-dev libjasper-dev libavcodec-dev libavformat-dev libswscale-dev libxine2-dev libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libv4l-dev libtbb-dev qtbase5-dev libfaac-dev libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev x264 v4l-utils unzip
1
2
3
4
cd opencv-3.1.0/
mkdir build 
cd build/
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_TBB=ON -D WITH_V4L=ON -D WITH_QT=ON -D WITH_OPENGL=ON -DCUDA_NVCC_FLAGS="-D_FORCE_INLINES" ..

ps:版本不支持高版本cuda

1
2
cd /home/cherrie/opencv-3.1.0/modules/cudalegacy/src/
sudo gedit graphcuts.cpp

把

1
#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER)

修改为

1
#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER) || (CUDART_VERSION >= 8000)
1
2
3
cd opencv-3.1.0/
cd build/
sudo make -j8
1
2
3
sudo make install
sudo /bin/bash -c 'echo "/usr/local/lib" > /etc/ld.so.conf.d/opencv.conf'
sudo ldconfig

anaconda安装