So guys who here has experience with both GeForce and Quadro cards from the same generation for KVM deployments?
I understand that when you use a Quadro you avoid error 43 and no longer have to hide the KVM and it allows the msrs to properly pass to the guest machine.
I would love to see some benchmarks with a Geforce and Quadro containing the same cuda core counts being from the same generation of cards.
Is there a performance benefit at all? Does this effect interrupts on the video card? What extra features are unlocked when the KVM no longer has to hide it selfs from the guest?
Does this effect certain cpu flags like constant_tsc nonstop_tsc from running correctly?
I know blocking msrs can effect lower level hardware features based on virtualization as the OS does not get the correct info it needs.
Anyone with experience on the issue would love to hear from you, as iam working on a great KVM guide that will be addressing what most are missing.