Windows Server 2022 VM Crash by install Driver for Tesla M60 (Hyper-V & VMWARE ESXI Host 8)

i installed a Virtual Machine with Server 2022 on a ESXI host and Win Server Host and redirect the GPU into the VM, but if i install the driver from NVIDIA the VM crashes… on both systems. this is so annoying.

what is the right way to install a M60 Tesla on a virtual machine?
Thank you.

A few more information would be helpful. What are you doing? vDGA (Passthrough)? When is it crashing? Which error?

hi sschaber,

it crashs on installation proccess by 30%, no error, VM crashs and restart.

on hyper-v its dda and on esxi both PCI device and dynmic pci.

this is error message from windows log
Der Computer wurde nach einem schwerwiegenden Fehler neu gestartet. Der Fehlercode war: 0x0000007e (0xffffffffc0000005, 0xfffff801e3233711, 0xffffae8a015d3fc8, 0xffffae8a015d37e0). Ein volles Abbild wurde gespeichert in: C:\Windows\MEMORY.DMP. Berichts-ID: 1bcdb36a-d8da-4560-8753-a57e2be7ab5c.

Have you tried another guest driver instead (vGPU 13.8) ?

from where?

and which cards can i install on hyper-v without extra costs for licencing?

You always need the vGPU driver for WDDM support with Datacenter GPUs. So you will also need to pay for vGPU licenses. vGPU 13.x is also available in the NV licensing portal.
Do you really expect me to suggest how to bypass vGPU licensing?

pls do not get me wrong, i only want to try to install a test station to test some things but nothing works. and direct passtrough is also a option for me but doesnt work because i prefer hyper-v.

where can i buy a vgpu licencing? i need one user. the complete system is not really simple

i cant find 13x, only 16.1 in the licecing portal

You should use the filter function in the portal.

It should be pretty easy to find the right vGPU version to download. As you already have portal access you already got a 90 days eval for vGPU.
Anyways, the driver install needs to work without the need for a vGPU license. It doesn’t matter if you use DDA or vDGA. Either way should work with M60 and the given guest driver.

oh thank you, i didnt see the filter function… thank you check it

this is crazy, driver can be installed, but now windows say:
Dieses Gerät wurde angehalten, weil es Fehler gemeldet hat. (Code 43)


Did you modify the high low MMIO space for DDA? This seems to be an issue with DDA:

i know this article but my actual VM is hosted on esxi 8, same drive issue

Where did you get this GPU from? Are you sure the GPU is already in graphics mode?

As this GPU is most likely 6-7 years old it might be that it is still in compute mode.

ebay… if i type this lspci –n | grep 10de
nothing happens.

i cant understand why the download links not posted in the FAQs… this is so complicated…

where can i get the GPU Switch software?

okay got it…

i think this can be the problem " Windows guest VMs are supported only on Q-series, B-series, and A-series NVIDIA vGPU types. "

GPU is now in Graphic mode, but nothing works and host crashs again (now i installed it on a server 2022 host system to check if the driver work. but driver for datacenter tesla m60 doesnt work… but why? Why have nvidia 2 drivers for one graphic card. i dont understand anything :(

i tested also a quadro p2000… same problem, driver not work. now i download driver and hope it works…

do i have first licence the server and then driver works?