caffe Linux 配置
1、安装CUDA，CUDA安装在系统目录下，用root账号安装,用root安装！！用root安装！！
CUDA从nvidia官网上下载，这里下载的是rpm包（cuda-repo-rhel7-7-5-local-7.5-18.x86_64.rpm）安装CUDA之前需要安装dkms
从 http://fedoraproject.org/wiki/EPEL 网址下载 epel-release-latest-7.noarch.rpm
用命令$rpm -ivh epel-release-latest-7.noarch.rpm 安装源包，
再更新yum源 $yum repolist
再安装 $yum install dkms libvdpau*.x86_64 freeglut*.x86_64
安装cuda：1、 rpm -i cuda-repo-rhel7-7-5-local-7.5-18.x86_64.rpm
2、 yum clean all
3、 yum install cuda
2、安装matlab
note:下边的$HOME=/home/yszhu
3、anaconda
bash Anaconda2-4.0.0-Linux-x86_64.sh
中间填写安装路径，安装在自己目录下的local/anaconda里
#添加环境变量
vi ~/.bashrc
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/home/yszhu/local/anaconda/lib"
export PATH="/home/yszhu/local/anaconda/bin:$PATH"
source ~/.bashrc
4、Install openblas
tar -zxvf OpenBLAS-0.2.18.tar.gz
cd OpenBLAS-0.2.18
mkdir -p $HOME/local/OpenBLAS
make -j8
make PREFIX=$HOME/local/OpenBLAS install
cd ../
rm -rf OpenBLAS-0.2.18
5、Install cmake [maybe not need if exist]
tar zxvf cmake-3.5.2.tar.gz
cd cmake-3.5.2
mkdir -p $HOME/local/cmake-3.5.2
./bootstrap --prefix=$HOME/local/cmake-3.5.2
make -j8
make install
#添加环境变量
vim ~/.bashrc
export PATH=/home/yszhu/local/cmake-3.5.2/bin:$PATH
source ~/.bashrc
cd ../
rm -rf cmake-3.5.2
6、Install Protobuf
#先安装c++版
tar zxvf protobuf-2.6.1.tar.gz
cd protobuf-2.6.1
./configure --prefix=$HOME/local
make -j8
make install
#继续安装Python版
cd python
python -V (Make sure you have Python 2.4 or newer)
python setup.py build （Build and run the tests）
python setup.py google_test
python setup.py install
cd ../../
rm -rf protobuf-2.6.1
7、Install snappy
tar zxvf snappy-1.1.3.tar.gz
cd snappy-1.1.3
./configure --prefix=$HOME/local
make -j8
make install
cd ../
rm -rf snappy-1.1.3
8、Install leveldb
tar zxvf leveldb-1.18.tar.gz
cd leveldb-1.18
make -j8
cp -av libleveldb.* $HOME/local/lib/
cp -av include/leveldb $HOME/local/include/
cd ../
rm -rf leveldb-1.18
9、Install OpenCV
tar zxvf opencv-2.4.8.tar.gz
cd opencv-2.4.8
mkdir release && cd release
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=$HOME/local -D BUILD_opencv_gpu=OFF -D CUDA_GENERATION=Kepler -D PYTHON_LIBRARY=$HOME/local/anaconda/bin ..
make -j8
make install
cd ../../
rm -rf opencv-2.4.8
10、Install Boost （安装旧版的没有bug）
tar zxvf boost_1_60_0.tar.gz
cd boost_1_60_0
./bootstrap.sh --prefix=$HOME/local
./b2 -j8
./b2 install
安装新版本的boost（boost_1_60_0）里面的gcc.h头文件定义存在bug，解决办法是安装完毕后：
vi $HOME/local/include/boost/config/compiler/gcc.hpp
定位到156行（#if defined(_GLIBCXX_USE_FLOAT128) && !defined(__STRICT_ANSI__)）
把156行修改为#if defined(_GLIBCXX_USE_FLOAT128) && !defined(__STRICT_ANSI__) && !defined(__CUDACC__)即可，不需要重新编译
安装旧版本的boost_1_55_0.tar不存在这个问题（建议安装旧版的）：
tar zxvf boost_1_55_0.tar.gz
cd boost_1_55_0
./bootstrap.sh --prefix=$HOME/local
./b2 -j8
./b2 install
cd ../
rm -rf boost_1_55_0
11、Install google-glog
tar zxvf glog-0.3.4.tar.gz
cd glog-0.3.4
./configure --prefix=$HOME/local
make -j8
make install
cd ../
rm -rf glog-0.3.4
12、Install gflags
tar zxvf gflags-2.1.2.tar.gz
cd gflags-2.1.2
mkdir build && cd build
CXXFLAGS="-fPIC" cmake -D CMAKE_INSTALL_PREFIX=$HOME/local ..
make -j8
make install
cd ../
rm -rf gflags-2.1.2
13、Install lmdb
tar zxvf lmdb-LMDB_0.9.18.tar.gz
cd lmdb-LMDB_0.9.18/libraries/liblmdb
make -j8
make prefix=$HOME/local install
cd ../../../
rm -rf lmdb-LMDB_0.9.18
14、Install hdf5
tar zxvf hdf5-1.8.14.tar.gz
cd hdf5-1.8.14
./configure --prefix=$HOME/local
make -j8
make check -j8 # run test suite.
make install
make check-install -j8 # verify installation.
cd ../
rm -rf hdf5-1.8.14
15、Install cuDNN
tar zxvf cudnn-7.0-linux-x64-v3.0-prod
cd cuda
cp -av lib64/* ~/local/lib
cp -av include/* ~/local/include
cd ../
rm -rf cuda
## add path to ~/.bashrc
#CUDA
export PATH=/ycs/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/ycs/local/cuda/lib64:$LD_LIBRARY_PATH
#for OpenBLAS
export LD_LIBRARY_PATH=/home/ycs/local/OpenBLAS/lib:$LD_LIBRARY_PATH
export OPENBLAS_NUM_THREADS=20
#for ~/local
export LD_LIBRARY_PATH=/home/ycs/local/lib:$LD_LIBRARY_PATH
export PATH=/home/ycs/local/bin:$PATH
source ~/.bashrc