Is it common to have more than one ID3D11Device and/or associated immediate ID3D11DeviceContext?

If I am correct a single display subsystem (GPU, video memory, etc.) is completely determined (from a 3D rendering perspective) by a

1. IDXGIAdapter (meta functionality facade);
2. ID3D11Device (resource creation facade);
3. ID3D11DeviceContext (pipeline facade).

So given that you want to use multiple display subsystems, you will have to handle multiple of these interfaces. A concrete example would be a graphics card dedicated to rendering and a separate graphics card dedicated to computation, or combining an integrated and dedicated graphics card. All such cases seem to me quite far fetched to justify support in a majority of games. So moving one abstraction level further downstream, should a game engine even consider multiple display systems (i.e. there is just one ID3D11Device and one immediate ID3D11DeviceContext)?

Your terminology seems to be coming from the D3D9 days, where a D3D9 adapter corresponded to a "display adapter." These days I'd consider a DXGI adapter and D3D11 device to correspond to a "render subsystem" rather than a "display subsystem." But that might just be nitpicking.

In D3D11, you're going to have a hard time using multiple adapters at the same time. Anything produced by a given GPU either needs to be presented by that GPU or funnelled back to the CPU, there's no way (at the moment at least) to marshal it directly to another GPU for processing.

In D3D12 on the other hand, cross-adapter resources are available, so you can actually use multiple GPUs in concert.

8 minutes ago, SoldierOfLight said:

so you can actually use multiple GPUs in concert

But do rendering engines exploit this?

I'm not aware of usage in a shipped engine, but there was a pretty cool demo a while back: https://blogs.msdn.microsoft.com/directx/2015/04/30/directx-12-multiadapter-lighting-up-dormant-silicon-and-making-it-work-for-you/

15 minutes ago, SoldierOfLight said:

I'm not aware of usage in a shipped engine, but there was a pretty cool demo a while back: https://blogs.msdn.microsoft.com/directx/2015/04/30/directx-12-multiadapter-lighting-up-dormant-silicon-and-making-it-work-for-you/

Thanks looks really impressive!

With AMD Crossfire / NVidia SLI, multiple physical GPUs might masquerade as a single D3D device. With 11, this is transparent magic performed by the driver (or if you want manual control, you can use the NVSDK/AGS extension libraries). With 12, devices now have a NodeMask parameter throughout the API, letting you specify work/resources for each physical GPU in the SLI/Crossfire system. 

17 hours ago, Hodgman said:

With AMD Crossfire / NVidia SLI, multiple physical GPUs might masquerade as a single D3D device. With 11, this is transparent magic performed by the driver (or if you want manual control, you can use the NVSDK/AGS extension libraries). With 12, devices now have a NodeMask parameter throughout the API, letting you specify work/resources for each physical GPU in the SLI/Crossfire system. 

This thus completely breaks the possibility of having singletons for them  (cfr. my other post) (Luckily, 90% of my methods passes these pointers around for use or storing [but not owning]. Though, I wrote some short hands to construct textures, buffers, shaders, etc. via constructors redirecting to the bigger constructors by using the singletons. So the whole I shot in my foot could be stitched )