First, version
Linux system: Ubuntu 14.04 (64-bit)
Graphics: Nvidia k20c
Cuda:cuda_7.5.18_linux.run
Cudnn:cudnn-7.0-linux-x64-v4.0-rc
Second, download
Ubuntu 14.04:http://www.ubuntu.com/download/desktop (64bit)
cuda7.5:https://developer.nvidia.com/cuda-downloads/, download the corresponding operating system and version cuda_7.5.18_linux.run, put it in the ~ root directory
CUDNN Download Address: HTTPS://DEVELOPER.NVIDIA.COM/CUDNN, need to register, and through the audit to download, download the corresponding file cudnn-7.0-linux-x64-v4.0-rc.tgz, put to ~ Root directory. Audit generally a two days, can not wait for the students to Http://pan.baidu.com/s/1bnOKBO download
Three, start the installation
1, the installation of Ubuntu, this article is not concerned about.
2. Disable Nouveau drive
Press CTRL+ALT+F1 to enter the command prompt to create a new blacklist file
# sudo vi/etc/modprobe.d/blacklist-nouveau.conf
Input
blacklist nouveauoptions nouveau modset=0
Save exit (: Wq)
And then execute
# sudo update-initramfs-u
Executive Lspci | grep nouveau See if there is content
# Lspci | grep nouveau
If there is no content, the description is disabled successfully, if there is content, reboot and then view
# sudo reboot
After restarting, enter the login screen, do not log into the desktop, directly press CTRL+ALT+F1 to enter the command prompt.
3, installation Cuda 7.5
Install some common software first, and note that the whole process requires networking
# sudo service lightdm stop # sudo apt-get install g++ # sudo apt-get install git # sudo apt-get install Freeglut3-dev
Next go to-install Cuda 7.5 According to the directory
# CD # sudo sh cude_7.5.18_linux.run
When installing, to let you look at a bunch of text (EULA), we press SPACEBAR to 100%, and then enter a bunch of accept,yes,yes or carriage return to install.
After the installation is complete, reboot and use LS to see if there are four or so folders that start with Nvidia
# ls/dev/nvidia*
If there is, the installation is successful, if not, may not succeed, need to unload load. The uninstall command is as follows:
# sudo/usr/local/cuda-7.5/bin/uninstall_cuda_7.5.pl # Sudo/usr/bin/nvidia-uninstall
If you are not sure if the installation is successful, please refer to the other tutorials to compile samples for testing.
Finally, configure the environment variables, we put directly in the system configuration file profile, first open the profile file
# sudo vi/etc/profile
Add two lines of code at the end, and if you don't edit with VI, please Baidu
Export path=/usr/local/cuda-7.5/bin: $PATHexport ld_library_path=/usr/local/cuda-7.5/lib64: $LD _library_ PATH
Save exit until Cuda 7.5 is installed.
4, installation Caffe
Download Caffe First
# sudo git clone https://github.com/BVLC/caffe.git
Then install a bunch of third-party libraries
#sudo apt-get install Libatlas-base-dev#sudo apt-get install Libprotobuf-dev#sudo apt-get install Libleveldb-dev#sudo apt-get install Libsnappy-dev#sudo apt-get install Libopencv-dev#sudo apt-get install Libboost-all-dev#sudo apt-get install Libhdf5-serial-dev#sudo apt-get install Libgflags-dev#sudo apt-get install Libgoogle-glog-dev#sudo apt-get install Liblmdb-dev#sudo apt-get install Protobuf-compiler
Next, install the OPENCV
# CD Caffe # sudo git clone https://github.com/jayambhia/Install-OpenCV # CD Install-opencv/ubuntu # sudo sh dependencies.sh # CD 2.4 # sudo sh opencv2_4_10.sh
Next, compile the Caffe
# CD ~/caffe # sudo cp Makefile.config.example makefile.config # make all
At this point, the Caffe installation is complete.
5. Test Caffe
Download Mnist data
# CD ~/caffe # sudo sh data/mnist/get_mnist.sh # sudo sh examples/mnist/create_mnist.sh
At runtime, if you have a GPU, you do not need to modify the configuration file, and if you do not have a GPU, you need to modify the configuration file Lenet_solver.prototxt
# sudo vi examples/mnist/lenet_solver.prototxt
Change the last line's Solver_mode:gpu to Solver_mode:cpu
Once configured, you are ready to run the
# sudo sh examples/mnist/train_lenet.sh
Note that when you run the Caffe program, you must be in the root directory of the Caffe, or you will get an error
6, install CUDNN to accelerate
Suppose you have already downloaded the cudnn-7.0-linux-x64-v4.0-rc.tgz to the ~ root directory.
#CD#sudo tar xvf cudnn-7.0-linux-x64-v4.0-rc.tgz#CD Cuda/include#sudo cp *.h/usr/local/include/#CD.. /lib64#sudo cp lib*/usr/local/lib/#Cd/usr/local/lib#sudo chmod +r libcudnn.so.4.0.4#sudo ln-sf libcudnn.so.4.0.4 libcudnn.so.4#sudo ln-sf libcudnn.so.4 libcudnn.so#sudo ldconfig
Remove the comments from the USE_CUDNN line in Makefile.config in the Caffek root directory and recompile.
# CD ~/caffe # sudo vi makefile.config
Remove the comment symbol # for the USE_CUDNN line, i.e. use_cudnn: = 1
and then recompile
# sudo make clean # sudo make all
Here, all installation is complete.
If you do not have a GPU, you cannot use CUDNN.
Learning in Ubuntu Caffe Series (1): Installation configuration Ubuntu14.04+cuda7.5+caffe+cudnn