Installing Azure Local and vGPU Configuration#
Note
Disclaimer
The content regarding GPU partitioning in this guide has been sourced from Azure documentation. For the most current and detailed information, please refer to the official Azure documentation.
Installing Azure Local#
To install Azure Local, follow the instructions from the Azure documentation.
GPU partitioning (or vGPU) allows multiple virtual machines (VMs) to share a single physical GPU. Each VM receives a dedicated portion of the GPU rather than access to the entire device. This feature utilizes the Single Root IO Virtualization (SR-IOV) interface, ensuring a hardware-backed security boundary and predictable performance for each VM. Secure partitioning prevents unauthorized access between VMs, making it ideal for workloads like virtual desktop infrastructure (VDI), AI, and ML inferencing. GPU partitioning can significantly reduce the total cost of ownership for your infrastructure.
Make sure to complete all the prerequisites before you begin to use the GPU partitioning feature.
Once you have your Azure Local cluster setup, it’s time to provision your GPU-enabled virtual machines.
NVIDIA vGPU Configuration#
NVIDIA vGPU Manager#
You need to install and configure the NVIDIA vGPU Manager. See Installing the NVIDIA Virtual GPU Manager for detailed instructions.
Selecting the Correct vGPU Profiles#
Select the appropriate vGPU profile based on your user’s needs and application requirements. More information can be found in the vGPU Sizing and Selection guides.
Creating the Virtual Machine#
Follow these instructions to create a virtual machine (VM) in Azure Local.
Adding a vGPU to the VM#
You add a vGPU to a Microsoft Azure Local VM by adding a GPU-P adapter to a VM. Follow these instructions to do so.
Installing the NVIDIA Graphics Driver#
After the VM is created and assigned a vGPU, install the NVIDIA vGPU Software Graphics Driver inside the VM. Instructions are available here.
Once completed, you can verify that the vGPU has been successfully added by running nvidia-smi inside PowerShell or the command prompt window in the VM.
Configuring Licensing#
The NVIDIA License System serves licenses to NVIDIA software products. To activate licensed functionalities, a licensed client leases a software license served over the network from an NVIDIA License System service instance. The NVIDIA License System Documentation explains how to install, configure, and manage licenses for virtual GPU software.
To configure a license client of the NVIDIA license system, start by generating a client configuration token, then follow these instructions.
To verify the license status, run nvidia-smi -q from the command line. You should see Licensed under Licensed Status in the vGPU Software Licensed Product.
You can also verify the license status through the NVIDIA Control Panel under Licensing → Manage License.
For troubleshooting issues, you can check the log files.
Managing the vGPU-enabled VM#
Migrating a VM Configured with vGPU#
Before migrating a VM configured with vGPU on Microsoft Azure Local, ensure the prerequisites are met.
To perform migration without failover clustering, see:
To perform migration with failover clustering, see Failover Clustering in Windows Server and Azure Local.
Note
Make sure to disable Trusted Platform Module (TPM) support in the virtual machine settings for any VM you want to migrate.