深度学习环境搭建关键步骤实测可用, 未禁nouvea: CUDA10.0+cudnn7.6.2 + tensorflow-gpu1.14.0 + ubuntu18.04 + pycharm2019.2

考虑到字体原因, 以下关键步骤均用英文. 经过实测, 整个过程未禁用 nouveau 

NVIDIA GTX 166ti + CUDA10.0+cudnn7.6.2 + tensorflow-gpu1.14.0 + ubuntu18.04 + pycharm2019.2

1,  Check the local environment under Ubuntu18.04  (在Ubuntu18.04下检查本机环境):

    https://ubuntu.com/download/desktop    # download ubuntu desktop version here if needed
    1>, Settting--> Details: Ubuntu 18.04.03
            #  should dispaly  "grapics: Intel" after the installation of Ubuntu.
    2>, Software & Updates: Additional Drivers:
           # should display two lines as following:  
           NVIDIA: nvidia-driver-430
           X.Org X server    # this line is selected

2,  download: cuda_10.0.130_410.48_linux.run; cudnn-10.0-linux-x64-v7.6.2.24 
      https://www.nvidia.cn/Download/index.aspx?lang=cn    # do NOT need download drivers
      https://developer.nvidia.com/cuda-10.0-download-archive
      https://developer.nvidia.com/rdp/cudnn-download       # need login first

3, in Software & Updates: Additional Drivers, change to Nvidia-driver, need reboot;
    then check: NVIDIA: "nvidia-driver-430"  is selected

4, install CUDA, do NOT install driver;

5, copy the cudnn:
    Copy the following files into the CUDA Toolkit directory, and change the file permissions.
    $ sudo cp cuda/include/cudnn.h /usr/local/cuda/include
    $ sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
    $ sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
    $ cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2

6, $ sudo gedit ~/.bashrc
    export CUDA_HOME=/usr/local/cuda
    export PATH=$PATH:$CUDA_HOME/bin
    export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
    $ source ~/.bashrc

7, $ sudo apt-get ./install python3-pip
    $ pip3 -V
    $ sudo apt-get remove python-pip  // uninstall pip for python2.7, don't do this~

8, pip3 install -r pip_install.txt, which contains:
    tensorflow-gpu
    numpy
    scipy
    matplotlib
    pillow

9, python3 ./gpuTest.py   ,which content is as following:
print("a test of gpu\n")
import tensorflow as tf
a = tf.constant([1.0,2.0,3.0],shape = [3], name='a')
b = tf.constant([1.0,2.0,3.0], shape = [3], name='b')
c = a +b
sess = tf.Session(config = tf.ConfigProto(log_device_placement =True))
cc = sess.run(c)
print(cc)

#should display as following:
add: (Add): /job:localhost/replica:0/task:0/device:GPU:0
a: (Const): /job:localhost/replica:0/task:0/device:GPU:0
b: (Const): /job:localhost/replica:0/task:0/device:GPU:0
[2. 4. 6.]

10, if can't use GPU in pycharm:
    pycharm --> Settings --> Project Interpreter --> Interpreter Paths --> add --> /usr/local/cuda-10.0/bin
    sudo ldconfig /usr/local/cuda-10.0/lib64  // for pycharm

11, if "could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR", using sess as:
    config = tf.ConfigProto()
    config.gpu_options.allow_growth=True
    sess = tf.Session(config=config)