Is it possible to use a nVidia GPU to render graphics if the monitor is connected to the motherboard video port

When I was learning the ropes on Slackware, there was this excellent newbie guide to installing, configuring and maintaining the distro. It has changed somewhat over the years, may not seem relevant to Debian & Co., but I still think that the SlackBook is a good piece of introductory material.

Not to pass any (imho due) judgement, but people seem to ignore the xorg manuals entirely. Maybe it's because they are made to believe that Xorg will "configure itself" as one may be accustomed with in Windows. This is not always the case and there is plenty of information out there on how to set up a dual monitor. I've answered at least two similar questions on this site about a very similar problem.

To the point then;

man xorg.conf

SERVERLAYOUT SECTION
(...)
Screen  screen-num "screen-id" position-information
(...)

The details are in the man page. What you want looks something like:

Section "ServerLayout"
    Identifier      "Three monitors"
    Screen  0       "Screen0" 0 0
    Screen  1       "Screen1" RightOf "Screen0"
    Screen  2       "Screen2" RightOf "Screen1"
EndSection

Checklist:

• Screens use correct device and monitor identifiers
• Screen has at least on Display sub section with a resolution and depth
• Drivers are loaded for the used devices

Notes:

Dual-head video cards (device section) may need special options such as bus and display mode. ATI cards used to "copy" the output to both heads in the past. This was preventable by passing certain options in the device section. Use the manual page of the xorg driver.

With recent versions of Xorg, xrandr can be used to configure and enable several monitors and to position them relative to each other. Having an xorg.config that configures several monitors is not really necessary, but may be more convenient.

Apparently the iGPU Multi-Monitor setting under advanced/system agent configuration/graphics configuration needs to be turned on. This is labelled for being for virtu but is necessary for the integrated GPU to be detectable by windows, even if you arn't running virtu.

Obviously the location of this may vary with different motherboard brands and the name of the setting may be different.

My guess is since lucid logix uses both GPUs, it tells the motherboard to simply run both - independent of whether you're running more than one monitor (with lucid outputting what you're rendering on the dedicated GPU to the integrated GPU), or using a multi-monitor set up taking advantage of both the IGPU and Dedicated GPU as independant outputs without lucid.

This might changed with later boards - but it was made possible by on die GPUs not using up PCIe lanes, and newer version of windows playing nice with more than one video card driver at a time.