nvidia_nightmare

tl;dr
use the package manager install (.deb or .rpm files, not .run file) 

remove stuff using 

sudo apt-get --purge remove  'cuda*'
sudo apt-get --purge -y remove 'nvidia*'
sudo apt-get --purge -y remove 'libnvidia*'

using 
jeremy@jeremy-Blade:~$ dpkg -l | grep -i nvidia

make sure everything ius iuninstalled 

uninstall steps to avoid 'existing runfile installation already found, it is strongly recommended to remove it'

sudo apt-get purge nvidia-current
sudo apt-get remove --purge nvidia-*

this may also hhelp
sudo /usr/local/cuda-X.Y/bin/uninstall_cuda_X.Y.pl
but in my case the cuda10.1 and 10.2 dir have no bin dir , and
$ sudo /usr/bin/nvidia-uninstall  doesnt exist

on running install i get 
jeremy@jeremy-Blade:~/Downloads$ sudo sh cuda_10.2.89_440.33.01_linux.run.1 
 Installation failed. See log at /var/log/cuda-installer.log for details.
jeremy@jeremy-Blade:~/Downloads$ more /var/log/cuda-installer.log
[INFO]: Driver not installed.
[INFO]: Checking compiler version...
[INFO]: gcc location: /usr/bin/gcc

[INFO]: gcc version: gcc version 7.4.0 (Ubuntu 7.4.0-1ubuntu1~18.04.1) 

[INFO]: Initializing menu
[INFO]: Setup complete
[INFO]: Components to install: 
[INFO]: Driver
[INFO]: 440.33.01
[INFO]: Executing NVIDIA-Linux-x86_64-440.33.01.run --ui=none --no-questions --accept-license --disable-nouveau --no-cc-versi
on-check --install-libglvnd  2>&1
[INFO]: Finished with code: 256
[ERROR]: Install of driver component failed.
[ERROR]: Install of 440.33.01 failed, quitting


dpkg -l | grep -i nvidia revealed

jeremy@jeremy-Blade:~$ dpkg -l | grep -i nvidia
rc  cuda-nvtx-10-1                                    10.1.243-1                                      amd64        NVIDIA Tools Extension
rc  cuda-nvtx-10-2                                    10.2.89-1                                       amd64        NVIDIA Tools Extension
rc  libnvidia-compute-415:amd64                       415.27-0ubuntu0~gpu18.04.2                      amd64        NVIDIA libcompute package
rc  libnvidia-compute-418:amd64                       418.87.01-0ubuntu1                              amd64        NVIDIA libcompute package
rc  libnvidia-compute-440:amd64                       440.33.01-0ubuntu1                              amd64        NVIDIA libcompute package


sudo apt-get remove --auto-remove nvidia-cuda-toolkit
didnt work - something else was using dpkg - so restarted and tried again 

sudo apt-get purge nvidia-cuda-toolkit or sudo apt-get purge --auto-remove nvidia-cuda-toolkit

Additionally, delete the /opt/cuda and ~/NVIDIA_GPU_Computing_SDK folders if they are present. and remove the export PATH=$PATH:/opt/cuda/bin and export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/cuda/lib:/opt/cuda/lib64 lines of the ~/.bash_profile file


this one removed a lot 

sudo apt-get --purge remove  'cuda*'
sudo apt-get --purge -y remove 'nvidia*'
sudo reboot
 
check  /var/log/cuda-installer.log

if that shows probs in driver then check 
/var/log/nvidia-installer.log

which shows
ERROR: You appear to be running an X server; please exit X before installing.  For further details, please see the section INSTALLING TH
E NVIDIA DRIVER in the README available on the Linux driver download page at www.nvidia.com.
so drop out to shell(alt-f4 on my ubuntu18.04)  and try 

sudo service lightdm stop 

unable to load nvidia.drm , also nouveau complaint. 

it turns out what i want is the .deb or package file and not the runfile (* It is recommended to use the distribution-specific packages, where possible. ) 



sudo nano /etc/profile

export PATH=/usr/local/cuda-10.2/bin:/usr/local/cuda-10.2/targets/x86_64-linux/lib/stubs${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.2/targets/x86_64-linux/lib/stubs${LD_LIBRARY_PATH:+:${LS_LIBRARY_PATH}}

 /usr/bin/nvidia-persistenced --verbose

check /var/crash, i found an err where module could not be made
ake[1]-***-no-rule-to-make-target-`modules'-stop-571

I tried 
jeremy@jeremy-Blade:~$ ln -s  /usr/local/cuda-10.2/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/x86_64-linux-gnu/libnvidia-ml.so
but still nvidia-smi still gives me 

NVIDIA-SMI couldn't find libnvidia-ml.so library in your system

and nvidia-settings gives me 
ERROR: Unable to load info from any available system

INFO 
jeremy@jeremy-Blade:~$ lspci |grep -i nvidia
01:00.0 VGA compatible controller: NVIDIA Corporation GP106M [GeForce GTX 1060 Mobile] (rev a1)

jeremy@jeremy-Blade:~$ uname -a 
Linux jeremy-Blade 5.3.0-26-generic #28~18.04.1-Ubuntu SMP Wed Dec 18 16:40:14 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

jeremy@jeremy-Blade:~$ lsb_release -a
No LSB modules are available.
Distributor ID:	Ubuntu
Description:	Ubuntu 18.04.3 LTS
Release:	18.04
Codename:	bionic

 1312  sudo ln -s  /usr/local/cuda-10.2/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/x86_64-linux-gnu/libnvidia-ml.so

and 
sudo ln -s  /usr/local/cuda-10.2/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/local/cuda/lib64/libnvidia-ml.so

didnt help

finally solved by install using pure package managers, no downloads 


CUDNN 

Check where your cuda installation is. For the installation from the repository it is /usr/lib/... and /usr/include. Otherwise, it will be /usr/local/cuda/ or /usr/local/cuda-<version>. You can check it with which nvcc or ldconfig -p | grep cuda
 Copy the files:

Repository installation:

$ cd folder/extracted/contents
$ sudo cp -P include/cudnn.h /usr/include
$ sudo cp -P lib64/libcudnn* /usr/lib/x86_64-linux-gnu/
$ sudo chmod a+r /usr/lib/x86_64-linux-gnu/libcudnn*

Glad it was of help to you