I have a PC with two braphics boards, Tesla K20c and GF119 (to which m display is connected).
I want to install CUDA and use it on the Tesla only.
For installing CUDA 10.2 on my CentOS 7 PC,
using RPM method (not run file) method as described here
https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#package-manager-installation
do I still need to disable the nouveau driver? (it is only mentioned for when one installs CUDA using the runFile installation
runFile installation