My company has 2 Grid K2, and we are giving them to 4 VMWare VMs, with passthrough, i.e. each VM with a whole GPU passthrough.
Last week I’ve created a new VM and assigned a spare GPU, but when I tested nvidia-smi -l I got:
WARNING: infoROM is corrupted at gpu 0000:13:00.0
This is the first time I receive it and I’ve been working with these K2 for 3 years. The other 3 VMs don’t report this warning either, with the same characteristics (same VMWare version, same OS, same NVIDIA driver).
Then, I shutdown the VM and assigned the GPU to a previous VM and it also show this message so it seems an issue with this specific GPU but not the other GPU inside the K2 graphic card.
I tried to find the meaning of this message, but I only found the description. Could it mean an issue in the hardware? Is there any other test I can run?