# Ubuntu16.04 install
**安裝步驟：**
1. 將光碟放入
2. 選擇中文繁體按安裝ubuntu
3. 選擇第三方軟體按繼續(等很久...)
4. 切割硬碟
5. 切割完開始安裝

**分割建議：**
1. 掛載點/；主分割區；安裝系統和軟件；大小為20~30G；分區格式為ext4；
2. 掛載點/home；邏輯分區；相當於“我的文檔”；大小為硬盤剩下的；分區格式ext4；
3. swap；邏輯分區；充當虛擬內存；大小等於內存大小2倍（4G）；分區格式為swap

# CUDA Install on Ubuntu 16.04

**Install nvcc & GPU driver**

$ sudo apt-get update
$ sudo apt install gcc-5
$ sudo apt install g++-5
$ sudo add-apt-repository ppa:graphics-drivers/ppa
$ sudo apt update
$ sudo apt-get install nvidia-390  (check your GPU driver version first)
$ sudo reboot

**重開後...check GPU**
check installation status

check GPU

**Install CUDA Newest(9.1)**
1. Go to website and download .deb
[Link](https://developer.nvidia.com/cuda-downloads)
2. 將下載好的.deb放到根目錄下(之後就不用在更換資料夾)
3. 重開 Ubuntu，並在登入畫面時按下 **Ctrl-Alt-F1** 進入命令列界面

4. 安裝 CUDA (直接照著download website的指令安裝)
5. 安裝完後重新開機

Reference: http://city.shaform.com/blog/2016/10/31/install-tensorflow-with-cuda.html

**Install CUDA 8.0** ([Link](https://developer.nvidia.com/cuda-80-ga2-download-archive))
步驟1~3同上接著以下安裝指令

$ sudo dpkg -i cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb
$ sudo apt-get update
$ sudo apt-get install cuda

接著調整PATH讓nvcc可以讀到
打開```/home/username/.bashrc```
在最後加上：

export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}}$ 
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

在termianl打上```source .bashrc```並reopen terminal
check CUDA version with ```nvcc --version```

# Install CUDNN on Ubuntu 16.04
1. 到 [website](https://developer.nvidia.com/rdp/cudnn-download) 下載 **cuDNN v5.1 Library for Linux** 移到home下
2.

    $ cd ~
    $ tar -zxf cudnn-8.0-linux-x64-v5.1.tgz
    $ cd cuda
    $ sudo cp lib64/* /usr/local/cuda/lib64/
    $ sudo cp include/* /usr/local/cuda/include/

reference: http://ywpkwon.github.io/post/ubuntu-setting/

**check Install version:**
1. check CUDA version:

2. check CUDNN version

# Install Dynet
先確認灌了CUDA 8.0 & CUDNN5.1 & python再開始灌dynet(網路上說cuda9會有問題)

$ sudo apt-get update
$ sudo apt-get install python-pip build-essential cmake mercurial

參考 dynet install python [website](http://dynet.readthedocs.io/en/latest/python.html) Manual Installation

**查看gcc version:**

$ gcc --version

**將gcc改成4.8** (不同CUDA搭配不同gcc version)

$ sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-4.8 10
$ sudo update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-4.8 10
$ sudo update-alternatives --config gcc
$ sudo update-alternatives --config g++ or (gcc-++)

check cmake 是否找到cudnn
在/dynet-base/dynet/build/裡打 ```cmake ..```
如果沒有cudnn的話：

others path setting:

# Anaconda Install on Ubuntu 16.04
1. 到 [website](https://docs.anaconda.com/anaconda/install/linux)下載 Anaconda3-5.0.1-Linux-x86_64.sh
2. Install python 3.6

    reopen terminal
3. 製作python2.7環境

# ubuntu 中文輸入法
參考： http://blog.xuite.net/yh96301/blog/342227672-Ubuntu+16.0.4%E6%96%B0%E5%A2%9E%E4%B8%AD%E6%96%87%E8%BC%B8%E5%85%A5%E6%B3%95-%E6%96%B0%E9%85%B7%E9%9F%B3%E8%BC%B8%E5%85%A5%E6%B3%95

# Sublime2 install

$ sudo add-apt-repository ppa:webupd8team/sublime-text-2
$ sudo apt-get update
$ sudo apt-get install sublime-text

## Reference
1. http://www.linuxandubuntu.com/home/how-to-install-latest-nvidia-drivers-in-linux
2. https://askubuntu.com/questions/885610/nvcc-version-command-says-nvcc-is-not-installed
##### tags: `ubuntu` `cuda` `dynet`