If you want to pass-through a consumer Nvidia card to a VM, you need to hide that you are running a VM from the guest OS, otherwise the Nvidia drivers will fail with some obscure errors.

https://github.com/sk1080/nvidia-kvm-patcher/blob/master/REA...

I've run into this before and gave up and just ran things on physical hw. It baffles me that we're running in to this crap in 2017.

Works fine for me. Now there is even software that doesn't require changing the HDMI input for this.

If you wanted to use a secondary nVidia card for PhysX, all was fine until nVidia changed their driver to disable PhysX when it detected an active AMD card.

NVIDIA GameWorks is probably the most cancerous thing to happen to the gaming industry. Partnering with UE4 to fuck games for AMD at the engine level :(

Here is another interesting read:

https://www.eevblog.com/forum/chat/hacking-nvidia-cards-into...

The difference between their "professional" GPUs and regular ones is literally a few sub-cent resistors.

Sometimes. The GeForce GTX 690 used the same chip as the Quadro K5000 (the GK104). More recent generations typically have different silicon (or bins with defective regions) in the pro and gaming lines. For instance, high end pascal gaming cards use GP104, but Tesla P100 uses the GP100.

For more details, see https://en.m.wikipedia.org/wiki/List_of_Nvidia_graphics_proc... (Notably the "code name" columns).

You used to be able to upgrade AMD CPUs with a pencil.

http://www.overclockersclub.com/guides/tbird/