After solved a login loop by purging and then reinstalling different nvidia drivers my titan x is now not used when performing deep learning, but instead my cpu. I uninstalled nvidia drivers but left cuda and cudnn, then reinstalled a different set of titan x drivers. Did I need to also uninstall cuda and cudnn as well, and then reinstall in order to be able to use the gpu again? Is there a more straightforward way than uninstalling everything and then reinstalling?
It’s hard to tell without more information what the issue at hand is. I can only assume you are talking about some sort of Linux distribution. If so, look at the installation guide for the precise steps to accomplish the installation:
I took a look through the installation guide and then decided to build a new virtual environment, set up with new settings and the latest tensorflow and now is recognising and using the gpu again. Thanks for the suggestion!