0

From the CUDA documents, I need to check whether my computer is capable by using the command: 

lspci | grep -i nvidia

but for my Thinkpad, this returns nothing. The document further says

If you do not see any settings, update the PCI hardware database that Linux maintains by entering update-pciids (generally found in /sbin) at the command line and rerun the previous lspci command.

I just don't understand what this means, who can tell me what to do next?

Ps, this is my computer:

OS: Ubuntu 18.04 LTS x86_64
Host: 4291RF4 ThinkPad X220
Kernel: 4.15.0-24-generic 
Uptime: 2 hours, 18 mins 
Packages: 3252 
Shell: bash 4.4.19 
Resolution: 1366x768 
DE: Xfce 
WM: Xfwm4 
WM Theme: Greybird 
Theme: Greybird [GTK2], Radiance [G 
Icons: Elementary-xfce-darker [GTK2 
Terminal: terminator 
CPU: Intel i3-2310M (4) @ 2.100GHz 
GPU: Intel Sandybridge Mobile 
Memory: 2265MiB / 7859MiB
graham
  • 10,436
John Paul Qiang Chen
  • 241
  • 1
    You don't have an Nvidia adapter. – Pilot6 Jul 09 '18 at 13:16
  • lspci doesn't "show nothing". You can run lspci and see what PCI devices are installed. The grep -i nvidia filters nvidia output. In your case you don't have any Nvidia device. – Pilot6 Jul 09 '18 at 13:27

2 Answers2

0

The message is telling to run the command update-pciids from the command line.

update-pciids(8)               The PCI Utilities              update-pciids(8)

NAME
       update-pciids - download new version of the PCI ID list


SYNOPSIS
       update-pciids [-q]


DESCRIPTION
       update-pciids  fetches the current version of the pci.ids file from the
       primary distribution site and installs it.

You will need to run it as root (sudo update-pciids) in order to write the updated file.

HOWEVER it most likely won't make any difference - as far as I know, the X220 has only Intel integrated graphics. Historically, Thinkpads with discrete graphics have been denoted by a "p" suffix.

steeldriver
  • 136,215
  • 21
  • 243
  • 336
0

CUDA is only available on Nvidia cards. Check list of CUDA capable graphics here. You have Intel GPU - try OpenCL instead of CUDA.

Bob
  • 2,533