Install Ubuntu16.04 LTS version

Install CUDA8.0 and cuDNN6

Thanks for this link: https://github.com/nicolaifsf/Installing-Tensorflow-with-GPU

1. Download CUDA Tookit8.0 runfile (local) https://developer.nvidia.com/cuda-80-ga2-download-archive

CUDA Tookit8.0

2. Download CUDNN from https://developer.nvidia.com/cudnn

cuDNN download

3. Download GTX1080Ti driver from http://www.nvidia.com/Download/index.aspx

GTX1080Ti driver

4. Run:

$ sudo apt-get install build-essential

5. Create the /etc/modprobe.d/blacklist-nouveau.conf file with

$ sudo vim /etc/modprobe.d/blacklist-nouveau.conf

In blacklist-nouveau.conf， input (you may need to install vim first)

blacklist nouveau
options nouveau modeset=0

Then,

$ sudo update-initramfs -u

6. Reboot your computer

$ reboot

7. On login screen, press ctrl + alt + f1 to open up login as tty1. You need to input your name and password in this step.
8. Navigate to the folder where you downloaded the files from steps 1, 2 and 3.
9. Run:

$ chmod +x cuda_8.0.61_375.26_linux.run
$ chmod +x NVIDIA-Linux-x86_64-378.13.run
$ sudo service lightdm stop
$ sudo apt-get remove --purge nvidia-*

10. Run:

$ sudo ./NVIDIA-Linux-x86_64-378.13.run --no-opengl-files

Just continue through the installation UNTILL IT ASKS YOU 'Would you like to run the nvidia-xconfig utility...' Hit NO for this step.

11. Run:

$ sudo ./cuda_8.0.61_375.26_linux.run --no-opengl-libs

Following the options below

Accept the EULA
No to Install NVIDIA Accelerated Graphics Driver for Linux-x86_64
Yes to Install the CUDA 8.0 Toolkit
Enter toolkit location to be  (default): /usr/local/cuda-8.0
Yes to install a symbolic link at /usr/local/cuda
Yes to install the CUDA 8.0 Samples
Default location for CUDA Samples location (or wherever you want them)

12. Run:

$ sudo modprobe nvidia

13. Set Environment path variables (add to .bashrc):

$ sudo vim ~/.bashrc

In the end of .bashrc, add

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

14. Run:

$ source ~/.bashrc
$ cat /proc/driver/nvidia/version
$ nvcc -V
$ sudo service lightdm start

15. Navigate to the CUDA samples we installed in step 11, then run:

$ make -j8

16. Navigate to: NVIDIA_CUDA-8.0_Samples/bin/x86_64/linux/release/ for the tests and run:

$ ./deviceQuery
$ ./deviceQuery

17. Reboot your computer

$ reboot

18. After rebooting, navigate to the folder where you downloaded the files from steps 1, 2 and 3 and run:

$ tar -xzvf cudnn-8.0-linux-x64-v6.0.tgz
$ sudo cp cuda/lib64/* /usr/local/cuda/lib64/
$ sudo cp cuda/include/cudnn.h /usr/local/cuda/include/

19. Run:

$ sudo apt-get install libcupti-dev

Congratualations! All CUDA dependencies for Tensorflow should now be installed.

Install TensorFlow r1.4

I used "native" pip method https://www.tensorflow.org/install/install_linux#InstallingNativePip and the python version is python3.5.