OS : ubuntu 14.04.5 desktop ( x64 )
Graphics: NVIDIA 1080Ti single card
motherboard no video interface (VGA DVI HDMI)

1 preparation

1> Download NVIDIA DRIVER

http://www.geforce.com/drivers
Manual search, the last item select recommended, the downloaded file is. RUN

2> Download CUDA

https://developer.nvidia.com/cuda-downloads
Select Runfile (LOCAL)
The file is very large. After downloading, remember to verify Checksum to avoid errors. Linux can use md5sum, win, please search FCIV

3> Download CUDNN

https://developer.nvidia.com/cudnn
At present, you need to log in to download. If you appear 403 when you download V6.0, please try after fan & qiang.

2 Check the hardware connection situation

lspci | grep VGA

If the connection is normal, you will seeNVIDIA Corporationwords

3 install NVIDIA Graphics Driver

nvidia-uninstall # Uninstall Driver
apt remove --purge nvidia*

Uninstall all package related to NVIDIA
If there is NOUVEAU, you need to turn it off, or uninstall it

Set boot to boot to the Console interface, edit/ETC/DEFAULT/GRUB
Keep the following, and update-grub

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash text"
GRUB_TERMINAL=console

Reboot and enter Console

lightdm stop  # Stop Light
chmod +x /path/to/NVIDIA-Linux-x86_64-xxx.run
/path/to/NVIDIA-Linux-x86_64-xxx.run  # Installation

There are prompts NVIDIA XCONFIG in the middle, because it will back up the configuration of the system itself, and it will recover when uninstalled this driver, so configure it for it yourself
After the installation is finished

[email protected]:/home/cnn# nvidia-smi
Fri Jul  7 10:28:16 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 381.22                 Driver Version: 381.22                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 108...  Off  | 0000:02:00.0     Off |                  N/A |
|  0%   39C    P0    55W / 250W |      0MiB / 11172MiB |      2%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

name has shown the model, NVIDIA -SMI -Q can see more details. Driver installation is complete

4 Install CUDA

execute the large RUN file downloaded earlier
First of all, it will display a long liveense, press Q, and then enter the access

1 Install NVIDIA Accelerated Graphics Driver for Linux-x86_64

Fill n
If this driver is installed, instead of installing files according to the separate Driver above 3, then executing NVIDIA-SMI and NVIDIA-SMI-Q will display the following content, which means that the graphics card is not recognized without identifying the graphics card.

nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 375.26                 Driver Version: 375.26                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  ERR!                Off  | 0000:02:00.0     Off |                  N/A |
|  0%   38C    P0    54W / 250W |      0MiB / 11171MiB |      3%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

nvidia-smi -q
Product Name                    : Unknown Error

2 Do you want to install the OpenGL libraries?

Fill n

3 Do you want to run nvidia-xconfig?

Fill y, the prompt will be recovered, don’t be false

4 Install the CUDA 8.0 Toolkit?

Enter Toolkit Location
 [ default is /usr/local/cuda-8.0 ]:
'Do you want to install a symbolic link at /usr/local/cuda?

If you have installed it and you do n’t uninstall it, you do n’t have to install it anymore. This LINK environment variable needs to be used, fill in y

5 Install the CUDA 8.0 Samples?

Enter CUDA Samples Location

SAMPLE can be used to detect whether CUDA is installed correctly.

After

, it is prompted to modify the path and ld_library_path

Driver:   Installed
Toolkit:  Installed in /usr/local/cuda-8.0
Samples:  Installed in /home/cnn/CUDA-samples, but missing recommended libraries

Please make sure that
 -   PATH includes /usr/local/cuda-8.0/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-8.0/lib64, or, add /usr/local/cuda-8.0/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run the uninstall script in /usr/local/cuda-8.0/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall

Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-8.0/doc/pdf for detailed information on setting up CUDA.

Logfile is /tmp/cuda_install_2602.log

Directly modify the /ETC /Profile to take effect on all users. Add the following content at the end of the file, and Source takes effect

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

Add link parameters

echo "/usr/local/cuda/lib64" > /etc/ld.so.conf.d/cuda.conf
ldconfig

5 Install CUDNN

Reference
https://developer.nvidia.com/compute/machine-learning/cudnn/secure/v6/prod/Doc/cudnn_install-txt
Installation separately
Runtime lib and DEV LIB

Reference

Ubuntu-Installation-CUDA7.0-Single Card Card-Super Detailed Tutorial
ubuntu 14.04 How to install the NVIDIA graphics card driver [reprinted]