Tesla M40 12GB passthrough issue

Hi Nvidia,

Could you please help me with issue.
I owned Tesla M40 12GB card, and want to use as encode GPU in VM on ESXi 6.7 in Windows 10.

When I enable passthrough in ESXi 6.7 U3, host needs to be rebooted.
After reboot GPU become disable in passthrough again, and VMware console asking to enable it again. And this is loop process which never ends, I have also FX3800; 4000 and 2000 Quadros, and ALL GPU’s works fine on same server with same ESXi 6.7 hypervisor.

Do you have any idea why I can’t passthrough M40 on host, I’m not using multi-GPUs on the same server, just tested with another GPU’s issue not with Motherboard…