If you’re running in Passthrough, there’s no point in installing the .vib in the Hypervisor as it won’t do anything. This will be why it’s working. If you were to try a vGPU Profile, you may have issues starting the VM.
As for SMI, this is expected, as you are not using any vGPUs so the Hypervisor can’t query them. Monitoring SMI within each VM will be your only option in this configuration.
Thank you Ben. My virtual machine (Ubuntu) is working with the nvidia driver vDGA. I don’t want to use the vGPU profile because it doesn’t support Linux (GRID K2).
But, is necessary that Xorg is running on the server?
Also, with vmware Which remote protocol should I use for vDGA? VMware Blast or PCoIP??
No, Xorg isn’t required for Passthrough. The Hypervisor isn’t doing anything with the GPU other than attaching it to your VMs.
Whether you’re using vGPU or Passthrough, the choice in protocol should be based on what you plan to deliver, not the GPU Profile. Ideally, your delivery technology will have a multi-codec protocol. This will give you the most flexibility.
Thanks for the answers, now, Which 3D renderer should use for Linux desktop pool with vDGA?
(Manage using vsphere Client, Automatic, Hardware, Software, vGPU or disabled)
"Do not change the Remote Display Protocol settings. These settings have no effect on Linux desktops. Also, the end user cannot choose the display protocol."