Step 1: Update and Upgrade your system

sudo apt-get update && sudo apt-get upgrade

Step 2: Verify You Have a CUDA-Capable GPU

lspci | grep -i nvidia

Step 3: Verify You Have a Supported Version of Linux

uname -m && cat /etc/*release

Step 4: Install Dependencies

确认安装python3.6

sudo apt-get install build-essential  && sudo apt-get install cmake git unzip zip && sudo add-apt-repository ppa:deadsnakes/ppa && sudo apt-get update && sudo apt-get install pylint

Step 5: Install linux kernel header

uname -r
You can get like “4.10.0-42-generic”. Note down linux kernel version. To install linux header supported by your linux kernel do following:

sudo apt-get install linux-image-extra-virtual
sudo apt-get install linux-source
sudo apt-get source linux-image-$(uname -r)
sudo apt-get install linux-headers-$(uname -r)

Step 6: Install NVIDIA CUDA 9.2

rm old version

sudo apt-get purge nvidia* && sudo apt-get autoremove && sudo apt-get autoclean && sudo rm -rf /usr/local/cuda*

本地运行：

rsync -avhP cuda_9.2.148_396.37_linux.run    [email protected]:/home/topppsen
rsync -avhP cudnn-9.2-linux-x64-v7.4.1.5.tgz  [email protected]:/home/topppsen

服务器运行

 sudo sh  cuda_9.2.148_396.37_linux.run --override --no-opengl-lib --kernel-source-path=/usr/src/linux-headers-$(uname -r)/

linux-headers-4.15.0-1024-gcp为uname -r所得
注意: 安装过程中，不要选择OpenGL，否则会出现，循环进入登录界面，本人选择安装 CUDA Samples（建议安装，待会会借助它查看是否安装成功，及显卡信息），并安装在 Documents 文件夹下。

重启服务器使cuda生效

添加环境变量：

echo 'export PATH=/usr/local/cuda-9.2/bin${PATH:+:${PATH}}' >> ~/.bashrc && echo 'export LD_LIBRARY_PATH=/usr/local/cuda-9.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}' >> ~/.bashrc && source ~/.bashrc

Check driver version probably Driver Version: 396.26
sudo ldconfig && nvidia-smi

终端进入你的 CUDA Samples安装目录，执行编译后运行 ./deviceQuery 输出显卡设备等信息。

cd NVIDIA_CUDA-9.2_Samples/1_Utilities/deviceQuery && make && ./deviceQuery

step 6: install CUDNN

将cudnn 安装文件上传到服务器，上面已经完成

进入服务器主目录解压
tar zxvf cudnn-9.2-linux-x64-v7.4.1.5.tgz

解压后，在你的目录下生成了一个“cuda”文件夹，对于cuDNN6.0的版本解压后生成“cudnn-8.0-linux-x64-v6.0”文件。使用如下命令copy，注意第二个有个-a参数，否则，拷贝过去的文件失去了链接。

copy the library files into CUDA's include and lib folders

sudo cp -R cuda/include/* /usr/local/cuda-9.2/include && sudo cp -R  cuda/lib64/* /usr/local/cuda-9.2/lib64

step 7: Tensorflow gpu install 见参考2 从step11开始

参考：
1.Ubuntu 16.04 LTS + CUDA8.0 + cudnn6.0
2.How to install Tensorflow GPU with CUDA 9.2 for python on Ubuntu
3.cuda_install_github

谷歌云GPU安装