0

I am looking to install Nvidia's proprietary drivers in order to be able to use CUDA for some Machine Learning.

But I am not able to select the proprietary driver option as the GPU model is not recognized in Software & Updates under the additional drivers tab: It only says Nvidia Corporation: Unknown

But if I lookup the GPU through a lspci command, it is detected correctly: It is indeed a GTX1660

Does someone know why that is the case?

limegreen
  • 1
  • Have you already tried to install latest NVIDIA drivers from webpage (https://www.nvidia.com/Download/driverResults.aspx/150803/en-us)? If not, you can give it a try. Please store your data before trying something with the driver, I ruined several times my Linux installation. For ML my personal best practice is to start from a fresh install and try to get the NVIDIA driver running. After driver installation verify that NVIDIA is detected (with 'nvidia-smi'). Check installation instruction on NVIDIA webpage (http://us.download.nvidia.com/XFree86/Linux-x86_64/435.21/README/installdriver.html) – Mr.Michael.Schulze Sep 11 '19 at 12:45
  • February 15 2019 — I think that GPU it's too new for Ubuntu. What version of Ubuntu do you use? – adasiko Sep 11 '19 at 12:55
  • Possible duplicate of How do I install the Nvidia drivers? – Kristopher Ives Sep 11 '19 at 14:27

1 Answers1

0

A manual install of the driver, indeed solved the issue. Well the GPU model still isn't recognized but I have a working and compatible driver. Thank you!

limegreen
  • 1