Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Azure Portal Deployment and Operations Guide with Switchless Networking
About using this guide Revision history Deployment Introduction Solution introduction Solution Deployment Deploy a cluster with Azure Stack HCI 23H2 operating system Update Azure Stack HCI clusters using SBE package GPU Integration
GPU Integration
Installing and configuring GPUs on AX platforms
AX-650 GPU configuration AX-750 GPU configuration AX-7525 GPU configuration GPU physical installation
GPU Discrete Device Assignment (DDA) on Azure Stack HCI
GPU partitioning
Host and guest driver download and install instructions
Registering a Partner Admin Link (PAL) for Azure related solutions Dell OpenManage Integration with Windows Admin Center Updates and maintenance References

GPU partitioning

Home > Integrated Products > Microsoft HCI Solutions from Dell Technologies > Guides 23H2 and later > Azure Portal Deployment and Operations Guide with Switchless Networking > GPU partitioning

GPU partitioning

  • Host and guest driver download and install instructions

    1. To download the drivers, go to https://nvid.nvidia.com/ and log in with your enterprise account.
    2. Click Nvidia licensing portal then on the left side click SOFTWARE DOWNLOADS.

      All the available software is listed for download.

    3. To locate Microsoft Azure stack HCI, apply filters PRODUCT FAMILY: VGPU and PLATFORM: Microsoft Azure Stack HCI.
      The following output appears:
      Figure 32. Driver downloads

      Driver downloads

      This package contains the host and drivers. Before proceeding to install the GPU-P drivers on the host, remove any Tesla drivers if previously installed.

    4. For more information on Installing the host driver, see Installing the NVIDIA Virtual GPU Manager for Microsoft Azure Stack of the NVIDIA Virtual GPU Software User Guide.
    5. After the host driver is installed on every cluster node, you can do GPU-P. For more information about how to perform GPU-P, see GPU partitioning.
      Note: It is highly recommended to use WAC and GPU extension.

    6. Once GPU partitions are assigned to the VM, install the GUEST driver. For more information on installing the guest driver, see Installing the NVIDIA vGPU SoftwareGraphics Driver on Windows of the NVIDIA Virtual GPU Software User Guide.
    7. After successful VM guest driver install, you must activate the guest VM with a NVIDIA license to start using the GPU-P. Details for doing so are outlined in Configuring a Licensed Client onWindows of the NVIDIA Virtual GPU Software User Guide.