install caffe

1)  download Universal-USB-Installer-1.9.6.0.exe, ubuntu-14.04.2-desktop-amd64.iso

mount usb flash to be an install device.

2) first time, using the usb, select UEFI. and then select the general settings.

3) when say install, select others to partition: select free space, '+' 

 20000 primary begining of the space , use as Ext4 mount point /

4000 logical beginning..as above use as swap area

500 as above use as Ext4 /boot

the rest as above /home

done ubuntu install

install cuda

1)verify if you have requirements

$ lspci | grep -i nvidia 

check if the type support cuda 

https://developer.nvidia.com/cuda-gpus

$ uname -m && cat /etc/*release

check if it is x86_64

$ gcc --version Verify the System Has gcc Installed

2) CTRL+ALT+F1

$ sudo stop lightdm

disable nouveau

$ cd /home/username
$ touch nvdia-graphics-drivers.conf 
$ sudo vi nvdia-graphics-drivers.conf
write in：blacklist nouveau

$ sudo cp nvdia-graphics-drivers.conf /etc/modprobe.d 

vi /etc/default/grub，add 

rdblacklist=nouveau nouveau.modeset=0 

reboot

$ sudo sh cuda_6.5.11_rc_linux_64.run

$ sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev

$ sudo start lightdm
 

 Environment Setup

$ export PATH=/usr/local/cuda-6.5/bin:$PATH
$ export LD_LIBRARY_PATH=/usr/local/cuda-6.5/lib64:$LD_LIBRARY_PATH

Verify the Installation

a. 验证驱动的版本，其实主要是保证驱动程序已经安装正常了

$ cat /proc/driver/nvidia/version
b. Compiling the Examples

$ nvcc -V

$ cd /home/username/NVIDIA_CUDA-6.5_Samples
$ make

$ cd /bin/x86_64/linux/release
$ ./deviceQuery

$ ./bandwidthTest

cuda done installation

caffe

1) blas  i used openblas easy fast need sudo apt-get install gFortran 

$ cd /etc/ld.so.conf.d

$ sudo touch openblas.conf

$ sudo vi openblas.conf

/opt/OpenBLASH/lib

$ sudo touch cuda.conf

$ sudo vi cuda.conf

/usr/local/cuda/lib64

/lib

 $ sudo ldconfig -v

2) opencv

some problems here with cuda

sudo apt-get install build-essential libgtk2.0-dev libjpeg-dev libtiff4-dev libjasper-dev libopenexr-dev cmake python-dev python-numpy python-tk libtbb-dev libeigen3-dev yasm libfaac-dev libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev libx264-dev libqt4-dev libqt4-opengl-dev sphinx-common texlive-latex-extra libv4l-dev libdc1394-22-dev libavcodec-dev libavformat-dev libswscale-dev default-jdk ant libvtk5-qt4-dev



cd ~

2wget http://sourceforge.net/projects/opencvlibrary/files/opencv-unix/2.4.9/opencv-2.4.9.zip

3unzip opencv-2.4.9.zip

4cd opencv-2.4.9

 

1mkdir build

2cd build

3cmake -D WITH_TBB=ON -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=ON -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D BUILD_EXAMPLES=ON -D WITH_QT=ON -D WITH_OPENGL=ON -D WITH_VTK=ON ..

the cmake has some problem with cuda.

try 

cmake command: cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_CUBLAS=ON -D WITH_CUFFT=ON -D WITH_EIGEN=ON -D BUILD_EXAMPLES=OFF -D BUILD_TESTS=OFF -D CUDA_ARCH_BIN="3.0" ..

or combine the cudapart with the first cmake option

replace http://code.opencv.org/projects/opencv/repository/revisions/feb74b125d7923c0bc11054b66863e1e9f753141/raw/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp 

with  the original hpp

make

make install

sudo gedit /etc/ld.so.conf.d/opencv.conf

/usr/local/lib

sudo ldconfig


sudo gedit /etc/bash.bashrc



1PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig

2export PKG_CONFIG_PATH

3) Other library

$ tar zxvf glog-0.3.3.tar.gz
$ ./ configure

$ make
$ sudo make install

$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libboost-all-dev libhdf5-serial-dev
  $ sudo apt-get install protobuf-c-compiler protobuf-compiler

3)

$ cp Makefile.config.example Makefile.config

“BLAS := open”，

change the blas lib and include /opt/openblas/lib include

$ make all
$ make test

$ make runtest 

4)

test

http://caffe.berkeleyvision.org/gathered/examples/mnist.html

cd $CAFFE_ROOT./data/mnist/get_mnist.sh./examples/mnist/create_mnist.sh

cd $CAFFE_ROOT./examples/mnist/train_lenet.sh

install matlab

matalb2014a part1 part2 crack in D 

mount -o loop,iocharset=gb2312 /tmp/download/Matlab_R14_Mac.Linux.Unix_CD1.iso /tmp/setup

unrar x -e file.part1.rar (will unrar 2 parts automatically if put in same folder)

sudo apt-get install rar unrar

sudo ./install

make matcaffe -j8

thanks to:

http://blog.csdn.net/u013476464/article/details/38071075

https://gist.github.com/bearpaw/c38ef18ec45ba6548ec0

http://www.samontab.com/web/2014/06/installing-opencv-2-4-9-in-ubuntu-14-04-lts/