How-to configure a Virtual GPU on an existing virtual machine

How-to configure a Virtual GPU on an existing virtual machine


Overview

This guide covers adding or removing an NVIDIA vGPU profile on an existing virtual machine in the ITS Private Cloud using the vss-cli command-line tool. By the end, your VM will have the requested vGPU profile attached (or removed) and be ready for driver installation.

The ITS Private Cloud vGPU offering is powered by NVIDIA L4 GPUs and supports both RTX Virtual Workstation (vWS) profiles for graphics and mixed graphics+compute workloads, and Virtual Compute Server (vCS) profiles for headless AI/ML and compute workloads. Available profiles are listed in the next section.

vGPU is only available on the FD1 and FD2 fault domains. If your VM is on a different domain, this guide includes the steps to migrate it.

After attaching a vGPU, the VM still needs the NVIDIA guest driver installed and a license token configured before the GPU is usable. See:How-to configure and licence a provisioned vGPU

vGPU Profiles

To tailor to various needs, we have made available a range of vGPU profiles as shown below:

Profile

Frame Buffer

Use Case

Profile

Frame Buffer

Use Case

nvidia_l4-6c

6GB

Mid-range compute, NLP inference, data science

nvidia_l4-12c

12GB

Large model inference, memory-intensive compute

nvidia_l4-24c

24GB

Full-GPU dedicated compute — AI training, large-scale inference

nvidia_l4-6q

6GB

High-end design, simulation, multi-monitor workflows

nvidia_l4-8q

8GB

High-res rendering, simulation, lighter AI inference

nvidia_l4-12q

12GB

Large model inference, scientific computing, hi-fi 3D

nvidia_l4-24q

24GB

Full-GPU dedicated workstation or compute VM


Step by Step - Add vGPU profile

Step 1 — Ensure the virtual machine is running on FD1 or FD2

Check the current FD

vss-cli compute vm get {vm-id-or-name} domain

If not in FD1 or FD2 use one of the commands below to move it (this may take minutes to hours depending on storage size and disk type)

vss-cli --wait compute vm set {vm-id-or-name} domain FD1

-OR-

vss-cli --wait compute vm set {vm-id-or-name} domain FD2

Step 2 — Shut down or power off your virtual machine

vss-cli --wait compute vm set {vm-id-or-name} state shutdown

Step 3 — Add the required vGPU profile

vss-cli --wait compute vm set {vm-id-or-name} gpu mk --profile nvidia_l4-6c

Step 4 — Confirm the profile was applied

vss-cli compute vm get {vm-id-or-name} gpu

Step 5 — Power on the virtual machine

vss-cli --wait compute vm set {vm-id-or-name} state on

Next Steps - Driver and License

Complete the GPU driver installation and licensing

If the vGPU creation fails, the reason is most likely because the profile is unavailable due to limited quantities.


Step by Step - Remove vGPU profile

Run vss-cli configure to configure your VSS credentials. If you have already done this, please go to next step.

Step 1 — Shut down or Power off your virtual machine

vss-cli --wait compute vm set {vm-id-or-name} state shutdown

Step 2 — Remove the GPU profile

vss-cli --wait compute vm set {vm-id-or-name} gpu rm

Step 3 — Confirm the GPU profile was removed

vss-cli compute vm get {vm-id-or-name} gpu

Step 4 — Power on the virtual machine

vss-cli --wait compute vm set {vm-id-or-name} state on

 Related articles

University of Toronto - Since 1827