Installing driver and required packages of libraries - CUDA, cudNN for NVIDIA GeForce RTX 2080

Hi,

In my new machine where Windows 10 was already installed, I have created two partitions on D drive. I have installed Ubuntu 16.04 on 2nd partition. Windows 10 already has NVIDIA Driver with the version that says 25.21.14.1967.

Now, I am trying to install NVIDIA driver for my Ubuntu 16.04 to support my current deep learning project. I am going to use TensorFlow 1.13.1 and according to post - https://devtalk.nvidia.com/default/topic/1047898/cuda-setup-and-installation/cuda-10-1-tensorflow-1-13/, cuda 10.0 version works fine.

I initially installed NVIDIA driver 430.14 using link - https://www.geforce.co.uk/drivers/results/147592. After installing it, clinfo command to check compatible platforms for GPU showed CUDA 10.2.120. What does that mean?

What I understood was that there is a CUDA with version 10.2.120 that has to be installed. I found none with this version. I thought installing the driver using .run file from the above link will do the trick of also installing CUDA library. When I did nvcc --version to check the version of CUDA there was none and it suggested me to run the command sudo apt install nvidia-cuda-toolkit which when I saw the trace of installation process seemed like it was installing another driver version. After finishing it, the previously installed NVIDIA driver got crashed.

Can somebody please help with how shall I go about installing the required for getting my TensorFlow running using this GPU?

Thanks in advance !!!

Regards,
Abinash Sinha