VIDEO_DXGKRNL_FATAL_ERROR in a Windows VM - workarounds

I’m finding using a particular tool called AquaSuite causes my Windows 10 VM to crash with VIDEO_DXGKRNL_FATAL_ERROR, but only after 10 minutes or so. It works fine otherwise.

AquaSuite is a client for setting up/controlling one of their USB devices that controls PC case fans/monitors temps.

When I Googled “VIDEO_DXGKRNL_FATAL_ERROR qemu”, there’s a clear pattern of monitoring tools. Couple of examples:

Corsair ICUE (similar software/USB device to AquaSuite)
MSI Afterburner.

In the Corsair case, all the responses I found were simply to stop using it. In the MSI Afterburner, turns out there was an option in the software to stop doing the low-level stuff.

I looked in the AquaSuite software options for similar - I turned off all GPU monitoring stuff I could find, but it’s still obviously doing it.

So the question is, is there something in Qemu I can do to stub a response? For instance, if the software insists on looking up some register that’s not there, rather than blowing up the OS, reply with duff data? How can I even debug what’s happening? There’s no host-side errors at all.

I thought about wrapping the AquaSuite software process, and preventing it event making the OS call, but to do that I first need to know what call it’s even making. However, the error is so severe there’s nothing in EventViewer, and no minidump.

Using AquaSuite in a VM - how can I prevent it attempting low-level hardware monitoring? - English forum - Aqua Computer Forum and onwards…look like there is a solution for the Aquasuite in a VM

Does Nvidia Inspector when used to change clocks also do this? I’ve used Afterburner, GPU Tweak II, and Aura lighting software on a Windows 8.1 VM without much issue on Fedora 27. Situation could be different with Windows 10.