If ESXi hosts in your vSphere with Tanzu environment have one or more NVIDIA GRID GPU graphics devices, you can configure VMs to use the NVIDIA GRID virtual GPU (vGPU) technology. You can also configure other PCI devices on an ESXi host to make them available to a VM in a passthrough mode.

NVIDIA GRID GPU
NVIDIA GRID GPU graphics devices are designed to optimize complex graphics operations and enable them to run at high performance without overloading the CPU. NVIDIA GRID vGPU provides unparalleled graphics performance, cost-effectiveness, and scalability by sharing a single physical GPU among multiple VMs as separate vGPU-enabled passthrough devices.

When you configure NVIDIA vGPU for a VM, you add a PCI device for vGPU to a VM class.

The following considerations apply when you use NVIDIA vGPU:
  • VMs with vGPU devices that are managed by VM Service are automatically powered off when an ESXi host enters maintenance mode. This might temporarily affect workloads running in the VMs. The VMs are automatically powered on after the host exists the maintenance mode.
Dynamic DirectPath I/O
Using Dynamic DirectPath I/O, the VM can directly access the physical PCI and PCIe devices connected to a host.

You can use Dynamic DirectPath I/O to assign multiple PCI passthrough devices to a VM. Each passthrough device can be specified by its PCI vendor and device identifier.

Prerequisites

  • Verify that the host machine is supported in the VMware Compatibility Guide, and check with the vendor to verify the host meets power and configuration requirements. Install a PCI device on the ESXi host.
  • To configure NVIDIA vGPU, follow these prerequisites:
    • Use vSphere version 7.0 Update 3 or later.
    • Configure ESXi host graphics settings with at least one device in Shared Direct mode. See Configuring Host Graphics.
    • Install NVIDIA vGPU software. NVIDIA provides a vGPU software package that includes the following components.

      For more information, see appropriate NVIDIA Virtual GPU Software documentation.

  • To configure Dynamic DirectPath I/O for PCI passthrough devices, follow these prerequisites:
  • Required privileges:
    • Namespaces.Modify cluster-wide configuration
    • Namespaces.Modify namespace configuration
    • Virtual Machine Classes.Manage Virtual Machine Classes

Procedure

  1. Add a PCI device to a VM class when you create or edit an existing VM class.
    Option Action
    Create a new VM class
    1. From the vSphere Client home menu, select Workload Management.
    2. Click the Services tab and click Manage on the VM Service pane.
    3. On the VM Service page, click VM Classes and click Create VM Class.
    4. On the Configuration page, specify the general VM class attributes. See Attributes of VM Classes in vSphere with Tanzu.

      Make sure that the memory resource reservation value is set to 100%.

    5. To add PCI devices, on the Configuration page, select Yes from the PCI Devices drop-down menu and click Next.
    Edit a VM class
    1. From the vSphere Client home menu, select Workload Management.
    2. Click the Services tab and click Manage on the VM Service pane.
    3. On the VM Service page, click VM Classes.
    4. In the existing VM class pane, click Manage and click Edit.

      Make sure that the memory resource reservation value is set to 100%.

    5. To add PCI devices, on the Configuration page, select Yes from the PCI Devices drop-down menu and click Next.
  2. On the PCI Devices page, expand the Add PCI Device menu, select the access type and other appropriate options, and click Next.
    Option Action
    NVIDIA GRID vGPU Specify the following options:
    • Model. Name of physical device. Select the device from the list of devices available on the host.
    • GPU Sharing. Indicates how physical GPU is shared across VMs. For example, Time Sharing.
    • GPU Mode. The GPU mode within a VM. For example, Compute is a configuration that is optimized for high-performance computing applications. While Workstation is used for graphics intensive workloads.
    • GPU Memory. Minimum GPU memory in GB per VM.
    • Number of vGPUs. Number of vGPU devices per VM.
    Dynamic DirectPath IO From the PCI Device list, select the PCI passthrough devices by their vendor, model name, or hardware label.
  3. On the Review and Confirm page, review the details and click Finish.

Results

A GPUs tag on the VM class pane indicates that the VM class is GPU-enabled.