You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I just found your repo and really appreciate the neat idea!
I'd like to know if there's any special considerations when using it with amd gpus as I can only see either intel integrated or --nvidia in the examples.
AMD is using open source drivers so a --device call should suffice AFAIK.
Another aspect I'm not sure of is that I'm also running an AMD cpu which has integrated graphics (that uses the same amdgpu driver). This means I have 2 cards exposed:
$ ls /dev/dri
by-path card0 card1 renderD128 renderD129
So I'm writting this issue to ask what would be the correct way to use it and if it would be possible to add an example to the documentation.
Cheers!
edit: a bit more info on the setup, both GPUs are exposed to X11 (the discrete card is actually used for offloading so maybe it would be interesting to add a way to change env var DRI_PRIME when using an amd dgpu?):
Hi,
I just found your repo and really appreciate the neat idea!
I'd like to know if there's any special considerations when using it with amd gpus as I can only see either intel integrated or
--nvidia
in the examples.AMD is using open source drivers so a
--device
call should suffice AFAIK.Another aspect I'm not sure of is that I'm also running an AMD cpu which has integrated graphics (that uses the same amdgpu driver). This means I have 2 cards exposed:
So I'm writting this issue to ask what would be the correct way to use it and if it would be possible to add an example to the documentation.
Cheers!
edit: a bit more info on the setup, both GPUs are exposed to X11 (the discrete card is actually used for offloading so maybe it would be interesting to add a way to change env var
DRI_PRIME
when using an amd dgpu?):The text was updated successfully, but these errors were encountered: