Hi everyone,
I remember, back when my old PC did not have an iGPU, I needed 2 dGPUs for passthrough to VM (1 GPU for host, 1 GPU for VM). Back then, I was it was important to choose a mainboard with great IOMMU groupings.
Nowadays, modern CPUs have iGPUs integrated. I am going to buy an AMD 9800x3d.
Idea:
- Use my iGPU for Linux host (probably Kubuntu)
- Use my dGPU for VM (Nvidia 4070 Super)
I want to keep my costs low and I do not want to invest too much money into a MB.
In this case, the dGPU is using the the standard GPU slot on the top of the mainboard. For my understanding, this slot is reserved for GPU usage only. There are not other devices associated to this IOMMU group.
Question 1:
If I understand it right, then I can buy just “any” mainboard, I am not dependent on choosing a premium mainboard with best-in-class IOMMU groupings?
Question 2:
Is it true that NVME / M2 slots come all “solo” without IOMMU pairing, so that I can just passthrough a whole SSD device to a VM? If I choose a MB with 4 NVME slots, then I can use any of them for passthrough?