Get NVIDIA vGPU performance and availability metrics – now in ControlUp

Like it or not, we can’t have a conversation these days about desktop virtualization without somebody mentioning GPU. The Concept of GPU use cases are proliferating far beyond graphical work forces to the task workers and beyond.

Today I’m happy to share we’ve started our first step in integrating NVIDIA Grid vGPU technology into the ControlUp Console.!

Live now, in the script based actions library, you’ll find two new scripts from yours truly as below:

Please note: these scripts are primarily supported on Driver version 369.71 on XenServer or ESXi. Earlier versions will work, but some detail is missing from the queries.

 

Get NVIDIA vGPU Performance Metrics:

This script based action will query for the availability of a vGPU in the system, then retrieve the following metrics:

  • Video Decoder / Encoder Usage (by percentage).
  • FrameBuffer Usage (by percentage).
  • GPU usage / load (by percentage).
  • GPU physical memory usage (by percentage).

An Example output can be found below:

 

Get NVIDIDA vGPU Details:

This script based action will query for the availability of a vGPU in the system, then retrieve the following pertinent information from the guest operating system:

  • Physical ID
  • vGPU Profile name
  • Architecture name
  • Core Count
  • Memory sizes (physical and virtual)
  • The current guest driver version
  • Licensing details (not applicable to early k1/k2 cards).

What’s next?

Go install the script based actions and query to your heart’s content! Expect to see these metrics in a later version of ControlUp coming soon ☺