Caffe installation and compilation process on INVIDIA+CPU core 2q9300 + ubuntu16.04.2

This article original, reproduced please indicate the source.

------------------------------------------------Split Line--------------------------------

Summary: The whole process is basically a fool mode, do not need to gcc4.9, do not need and gcc5.4 switch.

Hardware: Gtx650tiboost

Cool Core 2Q 93,004 Core

Motherboard looks like Gigabyte's an extremely old small board

Environment:

System: ubuntu16.04 86x64

gcc:5.4 86x64

Cuda8:cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb

CUDNN v5.0:cudnn-8.0-linux-x64-v5.0-ga.tgz here recommended to move it after decompression to cuda8, otherwise always hint can not find cudnn, very depressed. And if specified in the Cmake-gui, if the wrong is specified (I have tried to specify a static library, then there are many errors, I can not understand completely. Although it is trivial to say, but very can toss, also very pit. ), there may be a curious compilation error later, and there's no reason to find it. Therefore, it is recommended not to be stored separately for beginners.

Process:

1. Install the operating system, but do not update the system, do not install online, or network install third-party drivers, libraries, etc.

2. Sudo dpkg-i cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb

3. Sudo apt-get udpate

4. sudo apt-get install Cuda

5. sudo reboot

6. Nvidia-smi---Detect if the graphics driver is installed successfully

7. Lsmod | grep nvidia-detects if the graphics driver is installed successfully

If it is installed, continue below. If not, consider the first installation success again.

8.

sudo apt-get udpate-y && sudo apt-get upgrade-y

sudo apt-get install Libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev Protobuf-compiler

sudo apt-get install--no-install-recommends Libboost-all-dev

sudo apt-get install Libopenblas-dev liblapack-dev Libatlas-base-dev

sudo apt-get install Libgflags-dev libgoogle-glog-dev Liblmdb-dev

sudo apt-get install python Python-dev

9. sudo apt-get install cmake cmake-gui vim git wget hardinfo

Ten. Git clone https://github.com/BVLC/caffe.git

11. CD Caffe

mkdir Build

CD Build

CMake.

MAKE-J4 here is specifying four threads to compile (quad-core processor)

sudo make install

Complete.

--------------------------------------------------------------------------

Under normal circumstances, there will be no compilation errors.

But in the event of a case, consider whether the compiler is causing, or is not, dependent.

Caffe installation and compilation process on INVIDIA+CPU core 2q9300 + ubuntu16.04.2