Re:Fedora 27 released
by nbaksalyar on Tuesday November 14, 2017 @09:00PM
(#15700933)
> I experienced those crashes few times out of the blue
... which happens to be the exact opposite of what you want if you’re into deep learning.
Sigh.
Can one productively use a video card in a laptop or even in a high-end desktop for machine learning? At work we use a special workstation-type box with 4 NVidea cards. It is not connected to any display and we do not run any remote graphical sections on it that could have used graphical cards. Still it is slow at learning. If not the size of datasets (video in lossless compression), we would use cloud solutions.
Yeah, you can actually do it. Just run desktop in your integrated Intel GPU and CUDA in NVIDIA discrete GPU.
This is with Intel-only laptop (Dell XPS-13).
> Otherwise, we advise that you may wish to consider using the GNOME on Xorg session (see above) rather than the default Wayland session; this should at least prevent the crashes from ending your GNOME session when they occur. We do apologize for any inconvenience and/or lost data caused by such Shell crashes.
In Fedora 26 I experienced those crashes few times out of the blue. It is annoying to say the least.
[1] - https://fedoraproject.org/wiki/Common_F27_bugs#Wayland_issue...