Not really sure which channel to post this but we are delivering a CVAD Published Desktop with NVidia Tesla T4 cards using Hyper-V DDA for a CAD firm. overall its been ok, but randomly we get errors in the NVidia drivers and the GPU essentially disappears from the guest or just fails to start so it fails back to the standard display adapter. I am working with NVidia support on some of this related to drivers, etc, but wondered if anyone could help me with MIMO sizing? My current VM specs are 16 vCPU with 128GB RAM using XenApp and the MIMO is set to low 3GB and high 3GB. I have no idea on this other than what Microsoft documentation stated, but I am seeing different configurations around there. Anyone have a suggestion of what I should set that as at the Hyper-V VM level?