NVIDIA AI Enterprise User Guide

Documentation for administrators that explains how to install and configure NVIDIA AI Enterprise.

1. Introduction to NVIDIA AI Enterprise

NVIDIA® AI Enterprise is an end-to-end, cloud-native suite of AI and data analytics software, optimized so every organization can succeed with AI. It's certified to deploy anywhere—from the enterprise data center to the public cloud—and includes global enterprise support and training.

NVIDIA AI Enterprise includes key enabling technologies and software from NVIDIA for rapid deployment, management, and scaling of AI workloads in the modern hybrid cloud.

NVIDIA AI Enterprise enables the following:

  1. Leverage fully integrated, optimized, certified, and supported software from NVIDIA for AI workloads.

  2. Run NVIDIA AI frameworks and tools optimized for GPU acceleration, reducing deployment time and ensuring reliable performance.

  3. Deploy anywhere – including on popular data center platforms from VMware and Red Hat, mainstream NVIDIA-Certified Systems configured with or without GPUs, and on GPU-accelerated instances in the public cloud.

  4. Leverage the jointly certified NVIDIA and Red Hat solution to deploy and manage AI workloads in containers or VMs with optimized software.

  5. Scale out to multiple nodes, enabling even the largest deep learning training models to run on the VMware vSphere. Previously, scaling with bare metal performance in a fully virtualized environment was limited to a single node, limiting the complexity and size of AI workloads that could be supported.

  6. Run AI workloads at near bare-metal performance with new optimizations for GPU acceleration on vSphere, including support for the latest Ampere architecture including the NVIDIA A100. Additionally, technologies like GPUDirect Communications can now be supported on vSphere. This provides communication between GPU memory and storage across a cluster for improved performance.

1.1. NVIDIA AI Enterprise Software Architecture

The software in the NVIDIA AI Enterprise suite is organized into separate layers for infrastructure optimization software, cloud native deployment software, and AI and data science frameworks.

The content of these layers is as follows:

  • Infrastructure optimization software:
    • NVIDIA virtual GPU (vGPU) software
    • NVIDIA CUDA Toolkit
    • NVIDIA Magnum IO™ software stack for accelerated data centers
  • Cloud native deployment software:
    • NVIDIA GPU Operator
    • NVIDIA Network Operator
  • AI and data science frameworks:
    • TensorFlow
    • PyTorch
    • NVIDIA Triton Inference Server
    • NVIDIA TensorRT
    • RAPIDS

    The AI and data science frameworks are delivered as container images. Containerized software can be run directly with a tool such as Docker.

What Is Included?



Line diagram showing the organization of the software in NVIDIA AI Enterprise into multiple layers

The NVIDIA AI Enterprise suite includes frameworks that are broadly applicable and used across vertical industries such as manufacturing, logistics, financial services, retail, and healthcare.

NVIDIA AI Enterprise includes:

  1. TensorFlow and PyTorch for maching learning.

  2. NVIDIA TAO Toolkit for a faster, easier way to accelerate training and quickly create highly accurate and performant, domain-specific vision, and conversational AI models.

  3. NVIDIA Tensor RT, for GPU optimized deep learning inference and Triton Inference Server to deploy trained AI models at scale.

  4. Triton Inference Server supports all major frameworks, such as TensorFlow, TensorRT, PyTorch, MXNet, Python and more. Triton Inference Server also includes the RAPIDS FIL backend for the best inference performance for tree-based models on GPUs.

  5. NVIDIA RAPIDS, for end-to-end data science, machine learning and analytics pipeline.

  6. NVIDIA GPU and Network Operators, to deploy and manage NVIDIA GPU and Networking resources in Kubernetes.

  7. NVIDIA vGPU Software, to deploy vGPU on common data center platforms, including VMware and Red Hat.

1.2. Prerequisites for Using NVIDIA AI Enterprise

Before proceeding, ensure that these prerequisites are met:

  • You have a system that meets the requirements in NVIDIA AI Enterprise Release Notes.
  • One or more supported NVIDIA GPUs are installed in your system.
  • If you are using an NVIDIA A100 GPU, the following BIOS settings are enabled on your system:
    • Single Root I/O Virtualization (SR-IOV)
    • VT-d/IOMMU - Enabled
  • The following software is installed according to the instructions in the VMware documentation:
    • VMware vSphere Hypervisor ESXi
    • VMware vCenter Server
  • A VM to be enabled with NVIDIA vGPU is created with the virtual hardware resources in the following table.
    Resource Requirements
    vCPUs 16
    RAM 64 GB
    Storage 500 GB thin provisioned virtual disk
    NIC VMXNet3 NIC connected to network
  • A supported guest OS is installed in the VM.

For optimum performance, set options in your server configuration as follows:

  • Enable the following options:
    • Hyperthreading
    • Memory Mapped I/O above 4 GB (if applicable)
  • Set the Power Setting or System Profile option to High Performance.
  • If applicable, set CPU Performance to Enterprise or High Throughput.
Note: If NVIDIA card detection does not include all the installed GPUs, set this option to Enabled.

2. Installing and Configuring NVIDIA Virtual GPU Manager

The process for installing and configuring NVIDIA Virtual GPU Manager depends on the hypervisor that you are using. After you complete this process, you can install the display drivers for your guest OS and license any NVIDIA AI Enterprise licensed products that you are using.

2.1. About NVIDIA Virtual GPUs

2.1.1. NVIDIA vGPU Architecture

The high-level architecture of NVIDIA vGPU is illustrated in Figure 1. Under the control of the NVIDIA Virtual GPU Manager running under the hypervisor, NVIDIA physical GPUs are capable of supporting multiple virtual GPU devices (vGPUs) that can be assigned directly to guest VMs.

Guest VMs use NVIDIA vGPUs in the same manner as a physical GPU that has been passed through by the hypervisor: an NVIDIA driver loaded in the guest VM provides direct access to the GPU for performance-critical fast paths, and a paravirtualized interface to the NVIDIA Virtual GPU Manager is used for non-performant management operations.

Figure 1. NVIDIA vGPU System Architecture

Diagram showing the high-level architecture of NVIDIA vGPU

Each NVIDIA vGPU is analogous to a conventional GPU, having a fixed amount of GPU framebuffer, and one or more virtual display outputs or "heads". The vGPU’s framebuffer is allocated out of the physical GPU’s framebuffer at the time the vGPU is created, and the vGPU retains exclusive use of that framebuffer until it is destroyed.

Depending on the physical GPU, different types of vGPU can be created on the vGPU:

  • On all GPUs that support NVIDIA AI Enterprise, time-sliced vGPUs can be created.
  • Additionally, on GPUs that support the Multi-Instance GPU (MIG) feature, MIG-backed vGPUs can be created. The MIG feature is introduced on GPUs that are based on the NVIDIA Ampere GPU architecture.

2.1.1.1. Time-Sliced NVIDIA vGPU Internal Architecture

A time-sliced vGPU is a vGPU that resides on a physical GPU that is not partitioned into multiple GPU instances. All time-sliced vGPUs resident on a GPU share access to the GPU’s engines including the graphics (3D), video decode, and video encode engines.

In a time-sliced vGPU, processes that run on the vGPU are scheduled to run in series. Each vGPU waits while other processes run on other vGPUs. While processes are running on a vGPU, the vGPU has exclusive use of the GPU's engines. You can change the default scheduling behavior as explained in Changing Scheduling Behavior for Time-Sliced vGPUs.

Figure 2. Time-Sliced NVIDIA vGPU Internal Architecture

Diagram showing the internal architecture of a time-sliced NVIDIA vGPU

2.1.1.2. MIG-Backed NVIDIA vGPU Internal Architecture

A MIG-backed vGPU is a vGPU that resides on a GPU instance in a MIG-capable physical GPU. Each MIG-backed vGPU resident on a GPU has exclusive access to the GPU instance’s engines, including the compute and video decode engines.

In a MIG-backed vGPU, processes that run on the vGPU run in parallel with processes running on other vGPUs on the GPU. Process run on all vGPUs resident on a physical GPU simultaneously.

Figure 3. MIG-Backed NVIDIA vGPU Internal Architecture

Diagram showing the internal architecture of a MIG-backed NVIDIA vGPU

2.1.2. About Virtual GPU Types

The number of physical GPUs that a board has depends on the board. Each physical GPU can support several different types of virtual GPU (vGPU). vGPU types have a fixed amount of frame buffer, number of supported display heads, and maximum resolutions. They are grouped into different series according to the different classes of workload for which they are optimized. Each series is identified by the last letter of the vGPU type name.

Series Optimal Workload
C-series Compute-intensive server workloads, such as artificial intelligence (AI), deep learning, or high-performance computing (HPC)1, 2

The number after the board type in the vGPU type name denotes the amount of frame buffer that is allocated to a vGPU of that type. For example, a vGPU of type A16-4C is allocated 4096 Mbytes of frame buffer on an NVIDIA A16 board.

Due to their differing resource requirements, the maximum number of vGPUs that can be created simultaneously on a physical GPU varies according to the vGPU type. For example, an NVDIA A16 board can support up to 4 A16-4C vGPUs on each of its two physical GPUs, for a total of 16 vGPUs, but only 2 A16-8C vGPUs, for a total of 8 vGPUs.

When enabled, the frame-rate limiter (FRL) limits the maximum frame rate in frames per second (FPS) for C-series vGPUs to 60 FPS.

By default, the FRL is enabled for all GPUs. The FRL is disabled when the vGPU scheduling behavior is changed from the default best-effort scheduler on GPUs that support alternative vGPU schedulers. For details, see Changing Scheduling Behavior for Time-Sliced vGPUs. On vGPUs that use the best-effort scheduler, the FRL can be disabled as explained in the release notes for your chosen hypervisor at NVIDIA AI Enterprise Documentation.

Note:NVIDIA vGPU is a licensed product on all supported GPU boards. An NVIDIA AI Enterprise software license is required to enable all vGPU features within the guest VM.

For details of the virtual GPU types available from each supported GPU, see Virtual GPU Types for Supported GPUs.

2.1.3. Valid Virtual GPU Configurations on a Single GPU

Valid vGPU configurations on a single GPU depend on whether the vGPUs are time sliced or, on GPUs that support MIG, are MIG-backed.

2.1.3.1. Valid Time-Sliced Virtual GPU Configurations on a Single GPU

This release of NVIDIA vGPU supports time-sliced vGPUs with the same amount of frame buffer from different virtual GPU series on the same physical GPU. For example, A-series, B-series, C-series, and Q-series vGPUs with the same amount of frame buffer can reside on the same physical GPU simultaneously.

However, the requirement that all vGPUs have the same amount of frame buffer doesn’t extend across physical GPUs on the same card. Different physical GPUs on the same card may host virtual GPUs with different amounts of frame buffer at the same time, provided that the vGPU types on any one physical GPU all have the same amount of frame buffer.

For example, an NVIDIA A16 card has four physical GPUs, and can support several types of virtual GPU.
  • A configuration with a mixture of A16-4C vGPUs and A16-4Q vGPUs on GPU0 is valid.
  • A configuration with A16-16C vGPUs on GPU 0 and GPU 1, A16-8C vGPUs on GPU 2, and A16-4C vGPUs on GPU3 is valid.
  • A configuration with a mixture of A16-8C vGPUs and A16-4C vGPUs on GPU0 is invalid.

Not all hypervisors support time-sliced vGPUs with the same amount of frame buffer from different virtual GPU series on the same physical GPU. To determine if your chosen hypervisor supports this feature, consult the release notes for your hypervisor at NVIDIA AI Enterprise Documentation.

2.1.3.2. Valid MIG-Backed Virtual GPU Configurations on a Single GPU

This release of NVIDIA vGPU supports both homogeneous and mixed MIG-backed virtual GPUs based on the underlying GPU instance configuration.

For example, an NVIDIA A100 PCIe 40GB card has one physical GPU, and can support several types of virtual GPU. Figure 4 shows the following examples of valid homogeneous and mixed MIG-backed virtual GPU configurations on NVIDIA A100 PCIe 40GB.

  • A valid homogeneous configuration with 3 A100-2-10C vGPUs on 3 MIG.2g.10b GPU instances
  • A valid homogeneous configuration with 2 A100-3-20C vGPUs on 3 MIG.3g.20b GPU instances
  • A valid mixed configuration with 1 A100-4-20C vGPU on a MIG.4g.20b GPU instance, 1 A100-2-10C vGPU on a MIG.2.10b GPU instance, and 1 A100-1-5C vGPU on a MIG.1g.5b instance
Figure 4. Example MIG-Backed vGPU Configurations on NVIDIA A100 PCIe 40GB

Diagram showing examples of examples of valid homogeneous and mixed MIG-backed virtual GPU configurations on NVIDIA A100 PCIe 40GB.

2.2. Switching the Mode of a GPU that Supports Multiple Display Modes

Some GPUs support display-off and display-enabled modes but must be used in NVIDIA AI Enterprise deployments in display-off mode.

The GPUs listed in the following table support multiple display modes. As shown in the table, some GPUs are supplied from the factory in display-off mode, but other GPUs are supplied in a display-enabled mode.

GPU Mode as Supplied from the Factory
NVIDIA A40 Display-off
NVIDIA L40 Display-off
NVIDIA RTX 6000 Ada Display enabled
NVIDIA RTX A5000 Display enabled
NVIDIA RTX A5500 Display enabled
NVIDIA RTX A6000 Display enabled

A GPU that is supplied from the factory in display-off mode, such as the NVIDIA A40 GPU, might be in a display-enabled mode if its mode has previously been changed.

To change the mode of a GPU that supports multiple display modes, use the displaymodeselector tool, which you can request from the NVIDIA Display Mode Selector Tool page on the NVIDIA Developer website.

Note:

Only the following GPUs support the displaymodeselector tool:

  • NVIDIA A40
  • NVIDIA L40
  • NVIDIA RTX A5000
  • NVIDIA RTX 6000 Ada
  • NVIDIA RTX A5500
  • NVIDIA RTX A6000

Other GPUs that support NVIDIA AI Enterprise do not support the displaymodeselector tool and, unless otherwise stated, do not require display mode switching.

2.3. Downloading NVIDIA AI Enterprise

Before you begin, ensure that you have your order confirmation message and have created an NVIDIA Enterprise Account.
  1. Visit the NVIDIA Enterprise Application Hub by following the Login link in the instructions for using your NVIDIA Entitlement Certificate or when prompted after setting the password for your NVIDIA Enterprise Account.
  2. When prompted, provide your e-mail address and password, and click LOGIN.

    Screen capture showing the NVIDIA Enterprise Account page.

  3. On the NVIDIA APPLICATION HUB page that opens, click NVIDIA LICENSING PORTAL.

    The NVIDIA Licensing Portal dashboard page opens.



    Screen capture showing the NVIDIA Licensing Portal dashboard with no license servers.

    Note: Your entitlement might not appear on the NVIDIA Licensing Portal dashboard page until 24 business hours after you set your password during the initial registration process.
  4. In the NVIDIA Licensing Portal dashboard page opens, click the down arrow next to each entitlement listed to view details of the NVIDIA AI Enterprise that you purchased.

    Screen capture showing the NVIDIA Licensing Portal dashboard with all entitlements expanded.

  5. In the left navigation pane of the NVIDIA Licensing Portal dashboard, click SOFTWARE DOWNLOADS.
  6. On the Product Download page that opens, set the Product Family option to NVAIE and follow the Download link for NVIDIA AI Enterprise.
  7. When prompted to accept the license for the software that you are downloading, click AGREE & DOWNLOAD.
  8. When the browser asks what it should do with the file, select the option to save the file.

    After the download starts, a pop-up window opens for you to download any additional software that you might need for your NVIDIA AI Enterprise deployment.

  9. In the pop-up window, follow the links to download any additional software that you need for your NVIDIA AI Enterprise deployment.
    1. If you are using Delegated License Service (DLS) instances to serve licenses, follow the link to DLS 1.0 for your chosen hypervisor, for example, DLS 1.0 for VMware vSphere. For information about installing and configuring DLS instances, refer to NVIDIA License System User Guide.

2.4. Installing and Configuring the NVIDIA Virtual GPU Manager for Red Hat Enterprise Linux KVM

The following topics step you through the process of setting up a single Red Hat Enterprise Linux Kernel-based Virtual Machine (KVM) VM to use NVIDIA vGPU.

CAUTION:
Output from the VM console is not available for VMs that are running vGPU. Make sure that you have installed an alternate means of accessing the VM (such as a VNC server) before you configure vGPU.

Follow this sequence of instructions:

  1. Installing the NVIDIA Virtual GPU Manager for Red Hat Enterprise Linux KVM
  2. MIG-backed vGPUs only: Configuring a GPU for MIG-Backed vGPUs
  3. Getting the BDF and Domain of a GPU on a Linux with KVM Hypervisor
  4. Creating an NVIDIA vGPU on a Linux with KVM Hypervisor
  5. Adding One or More vGPUs to a Linux with KVM Hypervisor VM
  6. Setting vGPU Plugin Parameters on a Linux with KVM Hypervisor

After the process is complete, you can install the graphics driver for your guest OS and license any NVIDIA AI Enterprise licensed products that you are using.

2.4.1. Installing the NVIDIA Virtual GPU Manager for Red Hat Enterprise Linux KVM

The NVIDIA Virtual GPU Manager for Red Hat Enterprise Linux KVM is provided as a .rpm file.

CAUTION:
NVIDIA Virtual GPU Manager and guest VM drivers must be compatible. If you update vGPU Manager to a release that is incompatible with the guest VM drivers, guest VMs will boot with vGPU disabled until their guest vGPU driver is updated to a compatible version.

2.4.1.2. Verifying the Installation of the NVIDIA AI Enterprise for Red Hat Enterprise Linux KVM

After the Red Hat Enterprise Linux KVM server has rebooted, verify the installation of the NVIDIA AI Enterprise package for Red Hat Enterprise Linux KVM.
  1. Verify that the NVIDIA AI Enterprise package is installed and loaded correctly by checking for the VFIO drivers in the list of kernel loaded modules.
    # lsmod | grep vfio
    nvidia_vgpu_vfio       27099  0
    nvidia              12316924  1 nvidia_vgpu_vfio
    vfio_mdev              12841  0
    mdev                   20414  2 vfio_mdev,nvidia_vgpu_vfio
    vfio_iommu_type1       22342  0
    vfio                   32331  3 vfio_mdev,nvidia_vgpu_vfio,vfio_iommu_type1
    #
  2. Verify that the libvirtd service is active and running.
    # service libvirtd status
  3. Verify that the NVIDIA kernel driver can successfully communicate with the NVIDIA physical GPUs in your system by running the nvidia-smi command. The nvidia-smi command is described in more detail in NVIDIA System Management Interface nvidia-smi.
Running the nvidia-smi command should produce a listing of the GPUs in your platform.
# nvidia-smi
Fri Aug 17 18:46:50 2024
+------------------------------------------------------+
| NVIDIA-SMI 525.125.03   Driver Version: 525.125.03   |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M60           On   | 0000:85:00.0     Off |                  Off |
| N/A   23C    P8    23W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla M60           On   | 0000:86:00.0     Off |                  Off |
| N/A   29C    P8    23W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla P40           On   | 0000:87:00.0     Off |                  Off |
| N/A   21C    P8    18W / 250W |     53MiB / 24575MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
#

If nvidia-smi fails to run or doesn’t produce the expected output for all the NVIDIA GPUs in your system, see Troubleshooting for troubleshooting steps.

2.4.2. Getting the BDF and Domain of a GPU on a Linux with KVM Hypervisor

Sometimes when configuring a physical GPU for use with NVIDIA AI Enterprise, you must find out which directory in the sysfs file system represents the GPU. This directory is identified by the domain, bus, slot, and function of the GPU.

For more information about the directory in the sysfs file system that represents a physical GPU, see NVIDIA vGPU Information in the sysfs File System.

  1. Obtain the PCI device bus/device/function (BDF) of the physical GPU.
    # lspci | grep NVIDIA

    The NVIDIA GPUs listed in this example have the PCI device BDFs 06:00.0 and 07:00.0.

    # lspci | grep NVIDIA
    06:00.0 VGA compatible controller: NVIDIA Corporation GM204GL [Tesla M10] (rev a1)
    07:00.0 VGA compatible controller: NVIDIA Corporation GM204GL [Tesla M10] (rev a1)
  2. Obtain the full identifier of the GPU from its PCI device BDF.
    # virsh nodedev-list --cap pci| grep transformed-bdf
    transformed-bdf
    The PCI device BDF of the GPU with the colon and the period replaced with underscores, for example, 06_00_0.

    This example obtains the full identifier of the GPU with the PCI device BDF 06:00.0.

    # virsh nodedev-list --cap pci| grep 06_00_0
    pci_0000_06_00_0
  3. Obtain the domain, bus, slot, and function of the GPU from the full identifier of the GPU.
    virsh nodedev-dumpxml full-identifier| egrep 'domain|bus|slot|function'
    full-identifier
    The full identifier of the GPU that you obtained in the previous step, for example, pci_0000_06_00_0.

    This example obtains the domain, bus, slot, and function of the GPU with the PCI device BDF 06:00.0.

    # virsh nodedev-dumpxml pci_0000_06_00_0| egrep 'domain|bus|slot|function'
        <domain>0x0000</domain>
        <bus>0x06</bus>
        <slot>0x00</slot>
        <function>0x0</function>
          <address domain='0x0000' bus='0x06' slot='0x00' function='0x0'/>

2.4.3. Creating an NVIDIA vGPU on a Linux with KVM Hypervisor

For each vGPU that you want to create, perform this task in a Linux command shell on the a Linux with KVM hypervisor host.

Before you begin, ensure that you have the domain, bus, slot, and function of the GPU on which you are creating the vGPU. For instructions, see Getting the BDF and Domain of a GPU on a Linux with KVM Hypervisor.

How to create an NVIDIA vGPU on a Linux with KVM hypervisor depends on whether the NVIDIA vGPU supports single root I/O virtualization (SR-IOV). For details, refer to:

2.4.3.1. Creating a Legacy NVIDIA vGPU on a Linux with KVM Hypervisor

A legacy NVIDIA vGPU does not support SR-IOV.

  1. Change to the mdev_supported_types directory for the physical GPU.
    # cd /sys/class/mdev_bus/domain\:bus\:slot.function/mdev_supported_types/
    domain
    bus
    slot
    function
    The domain, bus, slot, and function of the GPU, without the 0x prefix.

    This example changes to the mdev_supported_types directory for the GPU with the domain 0000 and PCI device BDF 06:00.0.

    # cd /sys/bus/pci/devices/0000\:06\:00.0/mdev_supported_types/
  2. Find out which subdirectory of mdev_supported_types contains registration information for the vGPU type that you want to create.
    # grep -l "vgpu-type" nvidia-*/name
    vgpu-type
    The vGPU type, for example, M10-2Q.

    This example shows that the registration information for the M10-2Q vGPU type is contained in the nvidia-41 subdirectory of mdev_supported_types.

    # grep -l "M10-2Q" nvidia-*/name
    nvidia-41/name
  3. Confirm that you can create an instance of the vGPU type on the physical GPU.
    # cat subdirectory/available_instances
    subdirectory
    The subdirectory that you found in the previous step, for example, nvidia-41.

    The number of available instances must be at least 1. If the number is 0, either an instance of another vGPU type already exists on the physical GPU, or the maximum number of allowed instances has already been created.

    This example shows that four more instances of the M10-2Q vGPU type can be created on the physical GPU.

    # cat nvidia-41/available_instances
    4
  4. Generate a correctly formatted universally unique identifier (UUID) for the vGPU.
    # uuidgen
    aa618089-8b16-4d01-a136-25a0f3c73123
  5. Write the UUID that you obtained in the previous step to the create file in the registration information directory for the vGPU type that you want to create.
    # echo "uuid"> subdirectory/create
    uuid
    The UUID that you generated in the previous step, which will become the UUID of the vGPU that you want to create.
    subdirectory
    The registration information directory for the vGPU type that you want to create, for example, nvidia-41.

    This example creates an instance of the M10-2Q vGPU type with the UUID aa618089-8b16-4d01-a136-25a0f3c73123.

    # echo "aa618089-8b16-4d01-a136-25a0f3c73123" > nvidia-41/create

    An mdev device file for the vGPU is added to the parent physical device directory of the vGPU. The vGPU is identified by its UUID.

    The /sys/bus/mdev/devices/ directory contains a symbolic link to the mdev device file.

  6. Make the mdev device file that you created to represent the vGPU persistent.
    # mdevctl define --auto --uuid uuid
    uuid
    The UUID that you specified in the previous step for the vGPU that you are creating.
    Note: Not all Linux with KVM hypervisor releases include the mdevctl command. If your release does not include the mdevctl command, you can use standard features of the operating system to automate the re-creation of this device file when the host is booted. For example, you can write a custom script that is executed when the host is rebooted.
  7. Confirm that the vGPU was created.
    1. Confirm that the /sys/bus/mdev/devices/ directory contains the mdev device file for the vGPU.
      # ls -l /sys/bus/mdev/devices/
      total 0
      lrwxrwxrwx. 1 root root 0 Nov 24 13:33 aa618089-8b16-4d01-a136-25a0f3c73123 -> ../../../devices/pci0000:00/0000:00:03.0/0000:03:00.0/0000:04:09.0/0000:06:00.0/aa618089-8b16-4d01-a136-25a0f3c73123
    2. If your release includes the mdevctl command, list the active mediated devices on the hypervisor host.
      # mdevctl list
      aa618089-8b16-4d01-a136-25a0f3c73123 0000:06:00.0 nvidia-41

2.4.3.2. Creating an NVIDIA vGPU that Supports SR-IOV on a Linux with KVM Hypervisor

An NVIDIA vGPU that supports SR-IOV resides on a physical GPU that supports SR-IOV, such as a GPU based on the NVIDIA Ampere architecture.

  1. Enable the virtual functions for the physical GPU in the sysfs file system.
    Note:
    • Before performing this step, ensure that the GPU is not being used by any other processes, such as CUDA applications, monitoring applications, or the nvidia-smi command.
    • The virtual functions for the physical GPU in the sysfs file system are disabled after the hypervisor host is rebooted or if the driver is reloaded or upgraded.

    Use only the custom script sriov-manage provided by NVIDIA AI Enterprise for this purpose. Do not try to enable the virtual function for the GPU by any other means.

    # /usr/lib/nvidia/sriov-manage -e domain:bus:slot.function
    domain
    bus
    slot
    function
    The domain, bus, slot, and function of the GPU, without the 0x prefix.
    Note: Only one mdev device file can be created on a virtual function.

    This example enables the virtual functions for the GPU with the domain 00, bus 41, slot 0000, and function 0.

    # /usr/lib/nvidia/sriov-manage -e 00:41:0000.0
  2. Obtain the domain, bus, slot, and function of the available virtual functions on the GPU.
    # ls -l /sys/bus/pci/devices/domain\:bus\:slot.function/ | grep virtfn
    domain
    bus
    slot
    function
    The domain, bus, slot, and function of the GPU, without the 0x prefix.

    This example shows the output of this command for a physical GPU with slot 00, bus 41, domain 0000, and function 0.

    # ls -l /sys/bus/pci/devices/0000:41:00.0/ | grep virtfn
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn0 -> ../0000:41:00.4
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn1 -> ../0000:41:00.5
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn10 -> ../0000:41:01.6
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn11 -> ../0000:41:01.7
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn12 -> ../0000:41:02.0
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn13 -> ../0000:41:02.1
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn14 -> ../0000:41:02.2
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn15 -> ../0000:41:02.3
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn16 -> ../0000:41:02.4
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn17 -> ../0000:41:02.5
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn18 -> ../0000:41:02.6
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn19 -> ../0000:41:02.7
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn2 -> ../0000:41:00.6
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn20 -> ../0000:41:03.0
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn21 -> ../0000:41:03.1
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn22 -> ../0000:41:03.2
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn23 -> ../0000:41:03.3
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn24 -> ../0000:41:03.4
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn25 -> ../0000:41:03.5
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn26 -> ../0000:41:03.6
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn27 -> ../0000:41:03.7
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn28 -> ../0000:41:04.0
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn29 -> ../0000:41:04.1
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn3 -> ../0000:41:00.7
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn30 -> ../0000:41:04.2
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn31 -> ../0000:41:04.3
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn4 -> ../0000:41:01.0
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn5 -> ../0000:41:01.1
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn6 -> ../0000:41:01.2
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn7 -> ../0000:41:01.3
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn8 -> ../0000:41:01.4
    lrwxrwxrwx. 1 root root           0 Jul 16 04:42 virtfn9 -> ../0000:41:01.5
  3. Choose the available virtual function on which you want to create the vGPU and note its domain, bus, slot, and function.
  4. Change to the mdev_supported_types directory for the virtual function on which you want to create the vGPU.
    # cd /sys/class/mdev_bus/domain\:bus\:vf-slot.v-function/mdev_supported_types/
    domain
    bus
    The domain and bus of the GPU, without the 0x prefix.
    vf-slot
    v-function
    The slot and function of the virtual function.

    This example changes to the mdev_supported_types directory for the first virtual function (virtfn0) for the GPU with the domain 0000 and bus 41. The first virtual function (virtfn0) has slot 00 and function 4.

    # cd /sys/class/mdev_bus/0000\:41\:00.4/mdev_supported_types
  5. Find out which subdirectory of mdev_supported_types contains registration information for the vGPU type that you want to create.
    # grep -l "vgpu-type" nvidia-*/name
    vgpu-type
    The vGPU type, for example, A40-2Q.

    This example shows that the registration information for the A40-2Q vGPU type is contained in the nvidia-558 subdirectory of mdev_supported_types.

    # grep -l "A40-2Q" nvidia-*/name
    nvidia-558/name
  6. Confirm that you can create an instance of the vGPU type on the virtual function.
    # cat subdirectory/available_instances
    subdirectory
    The subdirectory that you found in the previous step, for example, nvidia-558.

    The number of available instances must be 1. If the number is 0, a vGPU has already been created on the virtual function. Only one instance of any vGPU type can be created on a virtual function.

    This example shows that an instance of the A40-2Q vGPU type can be created on the virtual function.

    # cat nvidia-558/available_instances
    1
  7. Generate a correctly formatted universally unique identifier (UUID) for the vGPU.
    # uuidgen
    aa618089-8b16-4d01-a136-25a0f3c73123
  8. Write the UUID that you obtained in the previous step to the create file in the registration information directory for the vGPU type that you want to create.
    # echo "uuid"> subdirectory/create
    uuid
    The UUID that you generated in the previous step, which will become the UUID of the vGPU that you want to create.
    subdirectory
    The registration information directory for the vGPU type that you want to create, for example, nvidia-558.

    This example creates an instance of the A40-2Q vGPU type with the UUID aa618089-8b16-4d01-a136-25a0f3c73123.

    # echo "aa618089-8b16-4d01-a136-25a0f3c73123" > nvidia-558/create

    An mdev device file for the vGPU is added to the parent virtual function directory of the vGPU. The vGPU is identified by its UUID.

  9. Time-sliced vGPUs only: Make the mdev device file that you created to represent the vGPU persistent.
    # mdevctl define --auto --uuid uuid
    uuid
    The UUID that you specified in the previous step for the vGPU that you are creating.
    Note:
    • If you are using a GPU that supports SR-IOV, the mdev device file persists after a host reboot only if you perform Step 1 before rebooting any VM that is configured with a vGPU on the GPU.
    • You cannot use the mdevctl command to make the mdev device file for a MIG-backed vGPU persistent. The mdev device file for a MIG-backed vGPU is not retained after the host is rebooted because MIG instances are no longer available.
    • Not all Linux with KVM hypervisor releases include the mdevctl command. If your release does not include the mdevctl command, you can use standard features of the operating system to automate the re-creation of this device file when the host is booted. For example, you can write a custom script that is executed when the host is rebooted.
  10. Confirm that the vGPU was created.
    1. Confirm that the /sys/bus/mdev/devices/ directory contains a symbolic link to the mdev device file.
      # ls -l /sys/bus/mdev/devices/
      total 0
      lrwxrwxrwx. 1 root root 0 Jul 16 05:57 aa618089-8b16-4d01-a136-25a0f3c73123 -> ../../../devices/pci0000:40/0000:40:01.1/0000:41:00.4/aa618089-8b16-4d01-a136-25a0f3c73123
    2. If your release includes the mdevctl command, list the active mediated devices on the hypervisor host.
      # mdevctl list
      aa618089-8b16-4d01-a136-25a0f3c73123 0000:06:00.0 nvidia-558

2.4.4. Adding One or More vGPUs to a Linux with KVM Hypervisor VM

To support applications and workloads that are compute or graphics intensive, you can add multiple vGPUs to a single VM.

Ensure that the following prerequisites are met:

You can add vGPUs to a Linux with KVM hypervisor VM by using any of the following tools:

  • The virsh command
  • The QEMU command line

After adding vGPUs to a Linux with KVM hypervisor VM, start the VM.

# virsh start vm-name
vm-name
The name of the VM that you added the vGPUs to.

After the VM has booted, install the NVIDIA AI Enterprise graphics driver as explained in Installing and Licensing NVIDIA AI Enterprise Components Natively.

2.4.4.1. Adding One or More vGPUs to a Linux with KVM Hypervisor VM by Using virsh

  1. In virsh, open for editing the XML file of the VM that you want to add the vGPU to.
    # virsh edit vm-name
    vm-name
    The name of the VM to that you want to add the vGPUs to.
  2. For each vGPU that you want to add to the VM, add a device entry in the form of an address element inside the source element to add the vGPU to the guest VM.
    <device>
    ...
      <hostdev mode='subsystem' type='mdev' model='vfio-pci'>
        <source>
          <address uuid='uuid'/>
        </source>
      </hostdev>
    </device>
    uuid
    The UUID that was assigned to the vGPU when the vGPU was created.

    This example adds a device entry for the vGPU with the UUID a618089-8b16-4d01-a136-25a0f3c73123.

    <device>
    ...
      <hostdev mode='subsystem' type='mdev' model='vfio-pci'>
        <source>
          <address uuid='a618089-8b16-4d01-a136-25a0f3c73123'/>
        </source>
      </hostdev>
    </device>

    This example adds device entries for two vGPUs with the following UUIDs:

    • c73f1fa6-489e-4834-9476-d70dabd98c40
    • 3b356d38-854e-48be-b376-00c72c7d119c
    <device>
    ...
      <hostdev mode='subsystem' type='mdev' model='vfio-pci'>
        <source>
          <address uuid='c73f1fa6-489e-4834-9476-d70dabd98c40'/>
        </source>
      </hostdev>
      <hostdev mode='subsystem' type='mdev' model='vfio-pci'>
        <source>
          <address uuid='3b356d38-854e-48be-b376-00c72c7d119c'/>
        </source>
      </hostdev>
    </device>
  3. Optional: Add a video element that contains a model element in which the type attribute is set to none.
    <video>
    <model type='none'/>
    </video>

    Adding this video element prevents the default video device that libvirt adds from being loaded into the VM. If you don't add this video element, you must configure the Xorg server or your remoting solution to load only the vGPU devices you added and not the default video device.

2.4.4.2. Adding One or More vGPUs to a Linux with KVM Hypervisor VM by Using the QEMU Command Line

Add the following options to the QEMU command line:
  • For each vGPU that you want to add to the VM, add one -device option in the following format:

    -device vfio-pci,sysfsdev=/sys/bus/mdev/devices/vgpu-uuid
    vgpu-uuid
    The UUID that was assigned to the vGPU when the vGPU was created.
  • Add a -uuid option to specify the VM as follows:

    -uuid vm-uuid
    vm-uuid
    The UUID that was assigned to the VM when the VM was created.

This example adds the vGPU with the UUID aa618089-8b16-4d01-a136-25a0f3c73123 to the VM with the UUID ebb10a6e-7ac9-49aa-af92-f56bb8c65893.

-device vfio-pci,sysfsdev=/sys/bus/mdev/devices/aa618089-8b16-4d01-a136-25a0f3c73123 \
-uuid ebb10a6e-7ac9-49aa-af92-f56bb8c65893

This example adds device entries for two vGPUs with the following UUIDs:

  • 676428a0-2445-499f-9bfd-65cd4a9bd18f
  • 6c5954b8-5bc1-4769-b820-8099fe50aaba

The entries are added to the VM with the UUID ec5e8ee0-657c-4db6-8775-da70e332c67e.

-device vfio-pci,sysfsdev=/sys/bus/mdev/devices/676428a0-2445-499f-9bfd-65cd4a9bd18f \
-device vfio-pci,sysfsdev=/sys/bus/mdev/devices/6c5954b8-5bc1-4769-b820-8099fe50aaba \
-uuid ec5e8ee0-657c-4db6-8775-da70e332c67e

2.4.5. Setting vGPU Plugin Parameters on a Linux with KVM Hypervisor

Plugin parameters for a vGPU control the behavior of the vGPU, such as the frame rate limiter (FRL) configuration in frames per second or whether console virtual network computing (VNC) for the vGPU is enabled. The VM to which the vGPU is assigned is started with these parameters. If parameters are set for multiple vGPUs assigned to the same VM, the VM is started with the parameters assigned to each vGPU.

For each vGPU for which you want to set plugin parameters, perform this task in a Linux command shell on the Linux with KVM hypervisor host.
  1. Change to the nvidia subdirectory of the mdev device directory that represents the vGPU.
    # cd /sys/bus/mdev/devices/uuid/nvidia
    uuid
    The UUID of the vGPU, for example, aa618089-8b16-4d01-a136-25a0f3c73123.
  2. Write the plugin parameters that you want to set to the vgpu_params file in the directory that you changed to in the previous step.
    # echo "plugin-config-params" > vgpu_params
    plugin-config-params
    A comma-separated list of parameter-value pairs, where each pair is of the form parameter-name=value.

    This example disables frame rate limiting and console VNC for a vGPU.

    # echo "frame_rate_limiter=0, disable_vnc=1" > vgpu_params

    This example enables unified memory for a vGPU.

    # echo "enable_uvm=1"  > vgpu_params

    This example enables NVIDIA CUDA Toolkit debuggers for a vGPU.

    # echo "enable_debugging=1"  > vgpu_params

    This example enables NVIDIA CUDA Toolkit profilers for a vGPU.

    # echo "enable_profiling=1"  > vgpu_params
To clear any vGPU plugin parameters that were set previously, write a space to the vgpu_params file for the vGPU.
# echo " " > vgpu_params

2.4.6. Deleting a vGPU on a Linux with KVM Hypervisor

For each vGPU that you want to delete, perform this task in a Linux command shell on the Linux with KVM hypervisor host.

Before you begin, ensure that the following prerequisites are met:

  1. Change to the mdev_supported_types directory for the physical GPU.
    # cd /sys/class/mdev_bus/domain\:bus\:slot.function/mdev_supported_types/
    domain
    bus
    slot
    function
    The domain, bus, slot, and function of the GPU, without the 0x prefix.

    This example changes to the mdev_supported_types directory for the GPU with the PCI device BDF 06:00.0.

    # cd /sys/bus/pci/devices/0000\:06\:00.0/mdev_supported_types/
  2. Change to the subdirectory of mdev_supported_types that contains registration information for the vGPU.
    # cd `find . -type d -name uuid`
    uuid
    The UUID of the vGPU, for example, aa618089-8b16-4d01-a136-25a0f3c73123.
  3. Write the value 1 to the remove file in the registration information directory for the vGPU that you want to delete.
    # echo "1" > remove

2.4.7. NVIDIA vGPU Information in the sysfs File System

Information about the NVIDIA vGPU types supported by each physical GPU in a Linux with KVM hypervisor host is stored in the sysfs file system.

All physical GPUs on the host are registered with the mdev kernel module. Information about the physical GPUs and the vGPU types that can be created on each physical GPU is stored in directories and files under the /sys/class/mdev_bus/ directory.

The sysfs directory for each physical GPU is at the following locations:

  • /sys/bus/pci/devices/
  • /sys/class/mdev_bus/

Both directories are a symbolic link to the real directory for PCI devices in the sysfs file system.

The organization the sysfs directory for each physical GPU is as follows:

/sys/class/mdev_bus/
           |-parent-physical-device
             |-mdev_supported_types
               |-nvidia-vgputype-id
                 |-available_instances
                 |-create
                 |-description
                 |-device_api
                 |-devices
                 |-name
parent-physical-device

Each physical GPU on the host is represented by a subdirectory of the /sys/class/mdev_bus/ directory.

The name of each subdirectory is as follows:

domain\:bus\:slot.function

domain, bus, slot, function are the domain, bus, slot, and function of the GPU, for example, 0000\:06\:00.0.

Each directory is a symbolic link to the real directory for PCI devices in the sysfs file system. For example:

# ll /sys/class/mdev_bus/
total 0
lrwxrwxrwx. 1 root root 0 Dec 12 03:20 0000:05:00.0 -> ../../devices/pci0000:00/0000:00:03.0/0000:03:00.0/0000:04:08.0/0000:05:00.0
lrwxrwxrwx. 1 root root 0 Dec 12 03:20 0000:06:00.0 -> ../../devices/pci0000:00/0000:00:03.0/0000:03:00.0/0000:04:09.0/0000:06:00.0
lrwxrwxrwx. 1 root root 0 Dec 12 03:20 0000:07:00.0 -> ../../devices/pci0000:00/0000:00:03.0/0000:03:00.0/0000:04:10.0/0000:07:00.0
lrwxrwxrwx. 1 root root 0 Dec 12 03:20 0000:08:00.0 -> ../../devices/pci0000:00/0000:00:03.0/0000:03:00.0/0000:04:11.0/0000:08:00.0
mdev_supported_types
A directory named mdev_supported_types is required under the sysfs directory for each physical GPU that will be configured with NVIDIA vGPU. How this directory is created for a GPU depends on whether the GPU supports SR-IOV.
  • For a GPU that does not support SR-IOV, this directory is created automatically after the Virtual GPU Manager is installed on the host and the host has been rebooted.
  • For a GPU that supports SR-IOV, such as a GPU based on the NVIDIA Ampere architecture, you must create this directory by enabling the virtual function for the GPU as explained in Creating an NVIDIA vGPU on a Linux with KVM Hypervisor. The mdev_supported_types directory itself is never visible on the physical function.
The mdev_supported_types directory contains a subdirectory for each vGPU type that the physical GPU supports. The name of each subdirectory is nvidia-vgputype-id, where vgputype-id is an unsigned integer serial number. For example:
# ll mdev_supported_types/
total 0
drwxr-xr-x 3 root root 0 Dec  6 01:37 nvidia-35
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-36
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-37
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-38
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-39
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-40
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-41
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-42
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-43
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-44
drwxr-xr-x 3 root root 0 Dec  5 10:43 nvidia-45
nvidia-vgputype-id
Each directory represents an individual vGPU type and contains the following files and directories:
available_instances
This file contains the number of instances of this vGPU type that can still be created. This file is updated any time a vGPU of this type is created on or removed from the physical GPU.
Note: When a time-sliced vGPU is created, the content of the available_instances for all other time-sliced vGPU types on the physical GPU is set to 0. This behavior enforces the requirement that all time-sliced vGPUs on a physical GPU must be of the same type. However, this requirement does not apply to MIG-backed vGPUs. Therefore, when a MIG-backed vGPU is created, available_instances for all other MIG-backed vGPU types on the physical GPU is not set to 0
create
This file is used for creating a vGPU instance. A vGPU instance is created by writing the UUID of the vGPU to this file. The file is write only.
description
This file contains the following details of the vGPU type:
  • The maximum number of virtual display heads that the vGPU type supports
  • The frame rate limiter (FRL) configuration in frames per second
  • The frame buffer size in Mbytes
  • The maximum resolution per display head
  • The maximum number of vGPU instances per physical GPU
For example:
# cat description
num_heads=4, frl_config=60, framebuffer=2048M, max_resolution=4096x2160, max_instance=4
device_api
This file contains the string vfio_pci to indicate that a vGPU is a PCI device.
devices
This directory contains all the mdev devices that are created for the vGPU type. For example:
# ll devices
total 0
lrwxrwxrwx 1 root root 0 Dec  6 01:52 aa618089-8b16-4d01-a136-25a0f3c73123 -> ../../../aa618089-8b16-4d01-a136-25a0f3c73123
name
This file contains the name of the vGPU type. For example:
# cat name
GRID M10-2Q

2.5. Installing and Configuring the NVIDIA Virtual GPU Manager for VMware vSphere

You can use the NVIDIA Virtual GPU Manager for VMware vSphere to set up a VMware vSphere VM to use NVIDIA vGPU or VMware vSGA. The vGPU Manager vSphere software components provide vSGA and vGPU functionality in a single component.

Note:

Some servers, for example, the Dell R740, do not configure SR-IOV capability if the SR-IOV SBIOS setting is disabled on the server. If you are using the Tesla T4 GPU with VMware vSphere on such a server, you must ensure that the SR-IOV SBIOS setting is enabled on the server.

However, with any server hardware, do not enable SR-IOV in VMware vCenter Server for the Tesla T4 GPU. If SR-IOV is enabled in VMware vCenter Server for T4, VMware vCenter Server lists the status of the GPU as needing a reboot. You can ignore this status message.

Requirements for Configuring NVIDIA vGPU in a DRS Cluster

You can configure a VM with NVIDIA vGPU on an ESXi host in a VMware Distributed Resource Scheduler (DRS) cluster. However, to ensure that the automation level of the cluster supports VMs configured with NVIDIA vGPU, you must set the automation level to Partially Automated or Manual.

For more information about these settings, see Edit Cluster Settings in the VMware documentation.

2.5.1. Installing the NVIDIA Virtual GPU Manager on VMware vSphere

To install the NVIDIA Virtual GPU Manager you need to access the ESXi host via the ESXi Shell or SSH. Refer to VMware’s documentation on how to enable ESXi Shell or SSH for an ESXi host.
Before you begin, ensure that the following prerequisites are met:
  • The ZIP archive that contains NVIDIA AI Enterprise has been downloaded from the NVIDIA Licensing Portal.
  • The software components for the NVIDIA Virtual GPU Manager have been extracted from the downloaded ZIP archive.
  1. Copy the NVIDIA Virtual GPU Manager component files to the ESXi host.
  2. Put the ESXi host into maintenance mode.
    $ esxcli system maintenanceMode set –-enable true
  3. Install the NVIDIA vGPU hypervisor host driver and the NVIDIA GPU Management daemon from their software component files.
    1. Run the esxcli command to install the NVIDIA vGPU hypervisor host driver from its software component file.
      $ esxcli software vib install -d /vmfs/volumes/datastore/host-driver-component.zip
    2. Run the esxcli command to install the NVIDIA GPU Management daemon from its software component file.
      $ esxcli software vib install -d /vmfs/volumes/datastore/gpu-management-daemon-component.zip
    datastore
    The name of the VMFS datastore to which you copied the software components.
    host-driver-component
    The name of the file that contains the NVIDIA vGPU hypervisor host driver in the form of a software component. Ensure that you specify the file that was extracted from the downloaded ZIP archive. For example, for VMware vSphere 7.0.2, host-driver-component is NVD-VMware-x86_64-525.125.03-1OEM.702.0.0.17630552-bundle-build-number.
    gpu-management-daemon-component
    The name of the file that contains the NVIDIA GPU Management daemon in the form of a software component. Ensure that you specify the file that was extracted from the downloaded ZIP archive. For example, for VMware vSphere 7.0.2, gpu-management-daemon-component is VMW-esx-7.0.2-nvd-gpu-mgmt-daemon-1.0-0.0.0001.
  4. Exit maintenance mode.
    $ esxcli system maintenanceMode set –-enable false
  5. Reboot the ESXi host.
    $ reboot

2.5.2. Updating the NVIDIA Virtual GPU Manager for VMware vSphere

Update the NVIDIA Virtual GPU Manager if you want to install a new version of NVIDIA Virtual GPU Manager on a system where an existing version is already installed.

To update the vGPU Manager VIB you need to access the ESXi host via the ESXi Shell or SSH. Refer to VMware’s documentation on how to enable ESXi Shell or SSH for an ESXi host.

Note: Before proceeding with the vGPU Manager update, make sure that all VMs are powered off and the ESXi host is placed in maintenance mode. Refer to VMware’s documentation on how to place an ESXi host in maintenance mode
  1. Stop the NVIDIA GPU Management Daemon.
    $ /etc/init.d/nvdGpuMgmtDaemon stop
  2. Update the NVIDIA vGPU hypervisor host driver and the NVIDIA GPU Management daemon.
    1. Run the esxcli command to update the NVIDIA vGPU hypervisor host driver from its software component file.
      $ esxcli software vib update -d /vmfs/volumes/datastore/host-driver-component.zip
    2. Run the esxcli command to update the NVIDIA GPU Management daemon from its software component file.
      $ esxcli software vib update -d /vmfs/volumes/datastore/gpu-management-daemon-component.zip
    datastore
    The name of the VMFS datastore to which you copied the software components.
    host-driver-component
    The name of the file that contains the NVIDIA vGPU hypervisor host driver in the form of a software component. Ensure that you specify the file that was extracted from the downloaded ZIP archive. For example, for VMware vSphere 7.0.2, host-driver-component is NVD-VMware-x86_64-525.125.03-1OEM.702.0.0.17630552-bundle-build-number.
    gpu-management-daemon-component
    The name of the file that contains the NVIDIA GPU Management daemon in the form of a software component. Ensure that you specify the file that was extracted from the downloaded ZIP archive. For example, for VMware vSphere 7.0.2, gpu-management-daemon-component is VMW-esx-7.0.2-nvd-gpu-mgmt-daemon-1.0-0.0.0001.
  3. Reboot the ESXi host and remove it from maintenance mode.

2.5.3. Verifying the Installation of the NVIDIA AI Enterprise Package for vSphere

After the ESXi host has rebooted, verify the installation of the NVIDIA AI Enterprise package for vSphere.
  1. Verify that the NVIDIA AI Enterprise package installed and loaded correctly by checking for the NVIDIA kernel driver in the list of kernel loaded modules.
    [root@esxi:~] vmkload_mod -l | grep nvidia
    nvidia                   5    8420
  2. If the NVIDIA driver is not listed in the output, check dmesg for any load-time errors reported by the driver.
  3. Verify that the NVIDIA GPU Management daemon has started.
    $ /etc/init.d/nvdGpuMgmtDaemon status
  4. Verify that the NVIDIA kernel driver can successfully communicate with the NVIDIA physical GPUs in your system by running the nvidia-smi command. The nvidia-smi command is described in more detail in NVIDIA System Management Interface nvidia-smi.
Running the nvidia-smi command should produce a listing of the GPUs in your platform.
[root@esxi:~] nvidia-smi
Fri Aug 17 17:56:22 2024
+------------------------------------------------------+
| NVIDIA-SMI 525.125.03   Driver Version: 525.125.03   |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M60           On   | 00000000:05:00.0 Off |                  Off |
| N/A   25C    P8    24W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla M60           On   | 00000000:06:00.0 Off |                  Off |
| N/A   24C    P8    24W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla M60           On   | 00000000:86:00.0 Off |                  Off |
| N/A   25C    P8    25W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Tesla M60           On   | 00000000:87:00.0 Off |                  Off |
| N/A   28C    P8    24W / 150W |     13MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
If nvidia-smi fails to report the expected output for all the NVIDIA GPUs in your system, see Troubleshooting for troubleshooting steps.

2.5.4. Managing the NVIDIA GPU Management Daemon for VMware vSphere

The NVIDIA GPU Management Daemon for VMware vSphere is a service that is controlled through scripts in the /etc/init.d directory. You can use these scripts to start the daemon, stop the daemon, and get its status.

  • To start the NVIDIA GPU Management Daemon, enter the following command:
    $ /etc/init.d/nvdGpuMgmtDaemon start
  • To stop the NVIDIA GPU Management Daemon, enter the following command:
    $ /etc/init.d/nvdGpuMgmtDaemon stop
  • To get the status of the NVIDIA GPU Management Daemon, enter the following command:
    $ /etc/init.d/nvdGpuMgmtDaemon status

2.5.5. Configuring VMware vMotion with vGPU for VMware vSphere

NVIDIA AI Enterprise supports vGPU migration, which includes VMware vMotion and suspend-resume, for VMs that are configured with vGPU. To enable VMware vMotion with vGPU, an advanced vCenter Server setting must be enabled. However, suspend-resume for VMs that are configured with vGPU is enabled by default.

Before configuring VMware vMotion with vGPU for an ESXi host, ensure that the current NVIDIA Virtual GPU Manager for VMware vSphere package is installed on the host.
  1. Log in to vCenter Server by using the vSphere Web Client.
  2. In the Hosts and Clusters view, select the vCenter Server instance.
    Note: Ensure that you select the vCenter Server instance, not the vCenter Server VM.
  3. Click the Configure tab.
  4. In the Settings section, select Advanced Settings and click Edit.
  5. In the Edit Advanced vCenter Server Settings window that opens, type vGPU in the search field.
  6. When the vgpu.hotmigrate.enabled setting appears, set the Enabled option and click OK.

    Screen capture showing the vgpu.hotmigrate.enabled setting in the Edit Advanced vCenter Server Settings window.

2.5.6. Changing the Default Graphics Type in VMware vSphere

The vGPU Manager VIB for VMware vSphere provides vSGA and vGPU functionality in a single VIB. After this VIB is installed, the default graphics type is Shared, which provides vSGA functionality. To enable vGPU support for VMs in VMware vSphere, you must change the default graphics type to Shared Direct.

If you do not change the default graphics type, VMs to which a vGPU is assigned fail to start and the following error message is displayed:

The amount of graphics resource available in the parent resource pool is insufficient for the operation.
Note: Change the default graphics type before configuring vGPU. Output from the VM console in the VMware vSphere Web Client is not available for VMs that are running vGPU.

Before changing the default graphics type, ensure that the ESXi host is running and that all VMs on the host are powered off.

  1. Log in to vCenter Server by using the vSphere Web Client.
  2. In the navigation tree, select your ESXi host and click the Configure tab.
  3. From the menu, choose Graphics and then click the Host Graphics tab.
  4. On the Host Graphics tab, click Edit.
    Figure 5. Shared default graphics type

    Screen capture of the Host Graphics tab in the VMware vCenter Web UI, showing the default graphics type as Shared

  5. In the Edit Host Graphics Settings dialog box that opens, select Shared Direct and click OK.
    Figure 6. Host graphics settings for vGPU

    Screen capture showing the Edit Host Graphics Settings dialog box in the VMware vCenter Web UI for changing the default graphics type

    Note: In this dialog box, you can also change the allocation scheme for vGPU-enabled VMs. For more information, see Modifying GPU Allocation Policy on VMware vSphere.

    After you click OK, the default graphics type changes to Shared Direct.

  6. Click the Graphics Devices tab to verify the configured type of each physical GPU on which you want to configure vGPU. The configured type of each physical GPU must be Shared Direct. For any physical GPU for which the configured type is Shared, change the configured type as follows:
    1. On the Graphics Devices tab, select the physical GPU and click the Edit icon.
      Figure 7. Shared graphics type

      Screen capture of the Graphics Devices tab in the VMware vCenter Web UI, showing the active type and configured type of two NVIDIA GPUs as Shared

    2. In the Edit Graphics Device Settings dialog box that opens, select Shared Direct and click OK.
      Figure 8. Graphics device settings for a physical GPU

      Screen capture showing the Edit Graphics Device Settings dialog box in the VMware vCenter Web UI for changing the graphics type of a physical GPU

  7. Restart the ESXi host or stop and restart nv-hostengine on the ESXi host.

    To stop and restart nv-hostengine, perform these steps:

    1. Stop nv-hostengine.
      [root@esxi:~] nv-hostengine -t
    2. Wait for 1 second to allow nv-hostengine to stop.
    3. Start nv-hostengine.
      [root@esxi:~] nv-hostengine -d
  8. In the Graphics Devices tab of the VMware vCenter Web UI, confirm that the active type and the configured type of each physical GPU are Shared Direct.
    Figure 9. Shared direct graphics type

    Screen capture of the Graphics Devices tab in the VMware vCenter Web UI, showing the active type and configured type of two NVIDIA GPUs as Shared Direct

After changing the default graphics type, configure vGPU as explained in Configuring a vSphere VM with NVIDIA vGPU.

See also the following topics in the VMware vSphere documentation:

2.5.7. Configuring a vSphere VM with NVIDIA vGPU

To support applications and workloads that are compute or graphics intensive, you can add multiple vGPUs to a single VM.

CAUTION:
Output from the VM console in the VMware vSphere Web Client is not available for VMs that are running vGPU. Make sure that you have installed an alternate means of accessing the VM (such as VMware Horizon or a VNC server) before you configure vGPU.

VM console in vSphere Web Client will become active again once the vGPU parameters are removed from the VM’s configuration.

How to configure a vSphere VM with a vGPU depends on your VMware vSphere version as explained in the following topics:

After you have configured a vSphere VM with a vGPU, start the VM. VM console in vSphere Web Client is not supported in this vGPU release. Therefore, use VMware Horizon or VNC to access the VM’s desktop.

After the VM has booted, install the NVIDIA AI Enterprise graphics driver as explained in Installing and Licensing NVIDIA AI Enterprise Software Components.

2.5.7.1. Configuring a vSphere 8 VM with NVIDIA vGPU

  1. Open the vCenter Web UI.
  2. In the vCenter Web UI, right-click the VM and choose Edit Settings.
  3. In the Edit Settings window that opens, configure the vGPUs that you want to add to the VM. Add each vGPU that you want to add to the VM as follows:
    1. From the ADD NEW DEVICE menu, choose PCI Device.
      Figure 10. Command for Adding a PCI Device

      Screen capture showing the PCI Device command on the ADD NEW DEVICE menu in the Edit Settings window in the VMware vCenter Web UI

    2. In the Device Selection window that opens, select the type of vGPU you want to configure and click SELECT.
      Figure 11. VM Device Selections for vGPU

      Screen capture showing VM settings for vGPU in the Device Selection window in the VMware vCenter Web UI

  4. Back in the Edit Settings window, click OK.

2.5.7.2. Configuring a vSphere 7 VM with NVIDIA vGPU

If you are adding multiple vGPUs to a single VM, perform this task for each vGPU that you want to add to the VM.
  1. Open the vCenter Web UI.
  2. In the vCenter Web UI, right-click the VM and choose Edit Settings.
  3. Click the Virtual Hardware tab.
  4. In the New device list, select Shared PCI Device and click Add. The PCI device field should be auto-populated with NVIDIA GRID vGPU.
    Figure 12. VM settings for vGPU

    Screen capture showing VM settings for vGPU in the Edit Settings window in the VMware vCenter Web UI

  5. From the GPU Profile drop-down menu, choose the type of vGPU you want to configure and click OK.
  6. Ensure that VMs running vGPU have all their memory reserved:
    1. Select Edit virtual machine settings from the vCenter Web UI.
    2. Expand the Memory section and click Reserve all guest memory (All locked).

2.5.8. Setting vGPU Plugin Parameters on VMware vSphere

Plugin parameters for a vGPU control the behavior of the vGPU, such as the frame rate limiter (FRL) configuration in frames per second or whether console virtual network computing (VNC) for the vGPU is enabled. The VM to which the vGPU is assigned is started with these parameters. If parameters are set for multiple vGPUs assigned to the same VM, the VM is started with the parameters assigned to each vGPU.

Ensure that the VM to which the vGPU is assigned is powered off.
For each vGPU for which you want to set plugin parameters, perform this task in the vSphere Client. vGPU plugin parameters are PCI pass through configuration parameters in advanced VM attributes.
  1. In the vSphere Client, browse to the VM to which the vGPU is assigned.
  2. Context-click the VM and choose Edit Settings.
  3. In the Edit Settings window, click the VM Options tab.
  4. From the Advanced drop-down list, select Edit Configuration.
  5. In the Configuration Parameters dialog box, click Add Row.
  6. In the Name field, type the parameter name pciPassthruvgpu-id.cfg.parameter, in the Value field type the parameter value, and click OK.
    vgpu-id
    A positive integer that identifies the vGPU assigned to a VM. For the first vGPU assigned to a VM, vgpu-id is 0. For example, if two vGPUs are assigned to a VM and you are setting a plugin parameter for both vGPUs, set the following parameters:
    • pciPassthru0.cfg.parameter
    • pciPassthru1.cfg.parameter
    parameter
    The name of the vGPU plugin parameter that you want to set. For example, the name of the vGPU plugin parameter for enabling unified memory is enable_uvm.
    To enable unified memory for two vGPUs that are assigned to a VM, set pciPassthru0.cfg.enable_uvm and pciPassthru1.cfg.enable_uvm to 1.

2.6. Configuring a GPU for MIG-Backed vGPUs

To support GPU instances with NVIDIA vGPU, a GPU must be configured with MIG mode enabled and GPU instances must be created and configured on the physical GPU. Optionally, you can create compute instances within the GPU instances. If you don't create compute instances within the GPU instances, they can be added later for individual vGPUs from within the guest VMs.

Ensure that the following prerequisites are met:

  • The NVIDIA Virtual GPU Manager is installed on the hypervisor host.
  • You have root user privileges on your hypervisor host machine.
  • You have determined which GPU instances correspond to the vGPU types of the MIG-backed vGPUs that you will create.

    To get this information, consult the table of MIG-backed vGPUs for your GPU in Virtual GPU Types for Supported GPUs.

  • The GPU is not being used by any other processes, such as CUDA applications, monitoring applications, or the nvidia-smi command.

To configure a GPU for MIG-backed vGPUs, follow these instructions:

  1. Enabling MIG Mode for a GPU

    Note: For VMware vSphere, only enabling MIG mode is required because VMware vSphere creates the GPU instances and, after the VM is booted and guest driver is installed, one compute instance is automatically created in the VM.
  2. Creating GPU Instances on a MIG-Enabled GPU

  3. Optional: Creating Compute Instances in a GPU instance

After configuring a GPU for MIG-backed vGPUs, create the vGPUs that you need and add them to their VMs.

2.6.1. Enabling MIG Mode for a GPU

Perform this task in your hypervisor command shell.

  1. Open a command shell as the root user on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Determine whether MIG mode is enabled. Use the nvidia-smi command for this purpose. By default, MIG mode is disabled.

    This example shows that MIG mode is disabled on GPU 0.

    Note: In the output from nvidia-smi, the NVIDIA A100 HGX 40GB GPU is referred to as A100-SXM4-40GB.
    $ nvidia-smi -i 0
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.03  Driver Version: 525.125.03   CUDA Version:  12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  A100-SXM4-40GB      On   | 00000000:36:00.0 Off |                    0 |
    | N/A   29C    P0    62W / 400W |      0MiB / 40537MiB |      6%      Default |
    |                               |                      |             Disabled |
    +-------------------------------+----------------------+----------------------+
  3. If MIG mode is disabled, enable it.
    $ nvidia-smi -i [gpu-ids] -mig 1
    gpu-ids
    A comma-separated list of GPU indexes, PCI bus IDs or UUIDs that specifies the GPUs on which you want to enable MIG mode. If gpu-ids is omitted, MIG mode is enabled on all GPUs on the system.

    This example enables MIG mode on GPU 0.

    $ nvidia-smi -i 0 -mig 1
    Enabled MIG Mode for GPU 00000000:36:00.0
    All done.
    Note: If the GPU is being used by another process, this command fails and displays a warning message that MIG mode for the GPU is in the pending enable state. In this situation, stop all processes that are using the GPU and retry the command.
  4. VMware vSphere ESXi only: Reboot the hypervisor host.
  5. Query the GPUs on which you enabled MIG mode to confirm that MIG mode is enabled.

    This example queries GPU 0 for the PCI bus ID and MIG mode in comma-separated values (CSV) format.

    $ nvidia-smi -i 0 --query-gpu=pci.bus_id,mig.mode.current --format=csv
    pci.bus_id, mig.mode.current
    00000000:36:00.0, Enabled

2.6.2. Creating GPU Instances on a MIG-Enabled GPU

Note: If you are using VMware vSphere, omit this task. VMware vSphere creates the GPU instances automatically.

Perform this task in your hypervisor command shell.

  1. If necessary, open a command shell as the root user on your hypervisor host machine.
  2. List the GPU instance profiles that are available on your GPU.

    You will need to specify the profiles by their IDs, not their names, when you create them.

    $ nvidia-smi mig -lgip
    +--------------------------------------------------------------------------+
    | GPU instance profiles:                                                   |
    | GPU   Name          ID    Instances   Memory     P2P    SM    DEC   ENC  |
    |                           Free/Total   GiB              CE    JPEG  OFA  |
    |==========================================================================|
    |   0  MIG 1g.5gb     19     7/7        4.95       No     14     0     0   |
    |                                                          1     0     0   |
    +--------------------------------------------------------------------------+
    |   0  MIG 2g.10gb    14     3/3        9.90       No     28     1     0   |
    |                                                          2     0     0   |
    +--------------------------------------------------------------------------+
    |   0  MIG 3g.20gb     9     2/2        19.79      No     42     2     0   |
    |                                                          3     0     0   |
    +--------------------------------------------------------------------------+
    |   0  MIG 4g.20gb     5     1/1        19.79      No     56     2     0   |
    |                                                          4     0     0   |
    +--------------------------------------------------------------------------+
    |   0  MIG 7g.40gb     0     1/1        39.59      No     98     5     0   |
    |                                                          7     1     1   |
    +--------------------------------------------------------------------------+
  3. Create the GPU instances that correspond to the vGPU types of the MIG-backed vGPUs that you will create.
    $ nvidia-smi mig -cgi gpu-instance-profile-ids
    gpu-instance-profile-ids
    A comma-separated list of GPU instance profile IDs that specifies the GPU instances that you want to create.

    This example creates two GPU instances of type 2g.10gb, which has profile ID 14.

    $ nvidia-smi mig -cgi 14,14
    Successfully created GPU instance ID  5 on GPU  2 using profile MIG 2g.10gb (ID 14)
    Successfully created GPU instance ID  3 on GPU  2 using profile MIG 2g.10gb (ID 14)

2.6.3. Optional: Creating Compute Instances in a GPU instance

Creating compute instances within GPU instances is optional. If you don't create compute instances within the GPU instances, they can be added later for individual vGPUs from within the guest VMs.
Note: If you are using VMware vSphere, omit this task. After the VM is booted and guest driver is installed, one compute instance is automatically created in the VM.

Perform this task in your hypervisor command shell.

  1. If necessary, open a command shell as the root user on your hypervisor host machine.
  2. List the available GPU instances.
    $ nvidia-smi mig -lgi
    +----------------------------------------------------+
    | GPU instances:                                     |
    | GPU   Name          Profile  Instance   Placement  |
    |                       ID       ID       Start:Size |
    |====================================================|
    |   2  MIG 2g.10gb      14        3          0:2     |
    +----------------------------------------------------+
    |   2  MIG 2g.10gb      14        5          4:2     |
    +----------------------------------------------------+
  3. Create the compute instances that you need within each GPU instance.
    $ nvidia-smi mig -cci -gi gpu-instance-ids
    gpu-instance-ids
    A comma-separated list of GPU instance IDs that specifies the GPU instances within which you want to create the compute instances.
    CAUTION:
    To avoid an inconsistent state between a guest VM and the hypervisor host, do not create compute instances from the hypervisor on a GPU instance on which an active guest VM is running. Instead, create the compute instances from within the guest VM as explained in Modifying a MIG-Backed vGPU's Configuration.

    This example creates a compute instance on each of GPU instances 3 and 5.

    $ nvidia-smi mig -cci -gi 3,5
    Successfully created compute instance on GPU  0 GPU instance ID  1 using profile ID  2
    Successfully created compute instance on GPU  0 GPU instance ID  2 using profile ID  2
  4. Verify that the compute instances were created within each GPU instance.
    $ nvidia-smi
    +-----------------------------------------------------------------------------+
    | MIG devices:                                                                |
    +------------------+----------------------+-----------+-----------------------+
    | GPU  GI  CI  MIG |         Memory-Usage |        Vol|         Shared        |
    |      ID  ID  Dev |           BAR1-Usage | SM     Unc| CE  ENC  DEC  OFA  JPG|
    |                  |                      |        ECC|                       |
    |==================+======================+===========+=======================|
    |  2    3   0   0  |      0MiB /  9984MiB | 28      0 |  2   0    1    0    0 |
    |                  |      0MiB / 16383MiB |           |                       |
    +------------------+----------------------+-----------+-----------------------+
    |  2    5   0   1  |      0MiB /  9984MiB | 28      0 |  2   0    1    0    0 |
    |                  |      0MiB / 16383MiB |           |                       |
    +------------------+----------------------+-----------+-----------------------+
    
    +-----------------------------------------------------------------------------+
    | Processes:                                                                  |
    |  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
    |        ID   ID                                                   Usage      |
    |=============================================================================|
    
    Note: Additional compute instances that have been created in a VM are destroyed when the VM is shut down or rebooted. After the shutdown or reboot, only one compute instance remains in the VM. This compute instance is created automatically after the NVIDIA AI Enterprise graphics driver is installed.

2.7. Disabling MIG Mode for One or More GPUs

If a GPU that you want to use for time-sliced vGPUs or GPU pass through has previously been configured for MIG-backed vGPUs, disable MIG mode on the GPU.

Ensure that the following prerequisites are met:

  • The NVIDIA Virtual GPU Manager is installed on the hypervisor host.
  • You have root user privileges on your hypervisor host machine.
  • The GPU is not being used by any other processes, such as CUDA applications, monitoring applications, or the nvidia-smi command.

Perform this task in your hypervisor command shell.

  1. Open a command shell as the root user on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Determine whether MIG mode is disabled. Use the nvidia-smi command for this purpose. By default, MIG mode is disabled, but might have previously been enabled.

    This example shows that MIG mode is enabled on GPU 0.

    Note: In the output from output from nvidia-smi, the NVIDIA A100 HGX 40GB GPU is referred to as A100-SXM4-40GB.
    $ nvidia-smi -i 0
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.03   Driver Version: 525.125.03  CUDA Version:  12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  A100-SXM4-40GB      Off  | 00000000:36:00.0 Off |                    0 |
    | N/A   29C    P0    62W / 400W |      0MiB / 40537MiB |      6%      Default |
    |                               |                      |              Enabled |
    +-------------------------------+----------------------+----------------------+
  3. If MIG mode is enabled, disable it.
    $ nvidia-smi -i [gpu-ids] -mig 0
    gpu-ids
    A comma-separated list of GPU indexes, PCI bus IDs or UUIDs that specifies the GPUs on which you want to disable MIG mode. If gpu-ids is omitted, MIG mode is disabled on all GPUs on the system.

    This example disables MIG mode on GPU 0.

    $ sudo nvidia-smi -i 0 -mig 0
    Disabled MIG Mode for GPU 00000000:36:00.0
    All done.
  4. Confirm that MIG mode was disabled. Use the nvidia-smi command for this purpose.

    This example shows that MIG mode is disabled on GPU 0.

    $ nvidia-smi -i 0
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.03   Driver Version: 525.125.03  CUDA Version:  12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  A100-SXM4-40GB      Off  | 00000000:36:00.0 Off |                    0 |
    | N/A   29C    P0    62W / 400W |      0MiB / 40537MiB |      6%      Default |
    |                               |                      |             Disabled |
    +-------------------------------+----------------------+----------------------+

2.8. Disabling and Enabling ECC Memory

Some GPUs that support NVIDIA AI Enterprise support error correcting code (ECC) memory with NVIDIA vGPU. ECC memory improves data integrity by detecting and handling double-bit errors. However, not all GPUs, vGPU types, and hypervisor software versions support ECC memory with NVIDIA vGPU.

On GPUs that support ECC memory with NVIDIA vGPU, ECC memory is supported with C-series and Q-series vGPUs, but not with A-series and B-series vGPUs. Although A-series and B-series vGPUs start on physical GPUs on which ECC memory is enabled, enabling ECC with vGPUs that do not support it might incur some costs.

On physical GPUs that do not have HBM2 memory, the amount of frame buffer that is usable by vGPUs is reduced. All types of vGPU are affected, not just vGPUs that support ECC memory.

The effects of enabling ECC memory on a physical GPU are as follows:

  • ECC memory is exposed as a feature on all supported vGPUs on the physical GPU.
  • In VMs that support ECC memory, ECC memory is enabled, with the option to disable ECC in the VM.
  • ECC memory can be enabled or disabled for individual VMs. Enabling or disabling ECC memory in a VM does not affect the amount of frame buffer that is usable by vGPUs.

GPUs based on the Pascal GPU architecture and later GPU architectures support ECC memory with NVIDIA vGPU. To determine whether ECC memory is enabled for a GPU, run nvidia-smi -q for the GPU.

Tesla M60 and M6 GPUs support ECC memory when used without GPU virtualization, but NVIDIA vGPU does not support ECC memory with these GPUs. In graphics mode, these GPUs are supplied with ECC memory disabled by default.

Some hypervisor software versions do not support ECC memory with NVIDIA vGPU.

If you are using a hypervisor software version or GPU that does not support ECC memory with NVIDIA vGPU and ECC memory is enabled, NVIDIA vGPU fails to start. In this situation, you must ensure that ECC memory is disabled on all GPUs if you are using NVIDIA vGPU.

2.8.1. Disabling ECC Memory

If ECC memory is unsuitable for your workloads but is enabled on your GPUs, disable it. You must also ensure that ECC memory is disabled on all GPUs if you are using NVIDIA vGPU with a hypervisor software version or a GPU that does not support ECC memory with NVIDIA vGPU. If your hypervisor software version or GPU does not support ECC memory and ECC memory is enabled, NVIDIA vGPU fails to start.

Where to perform this task depends on whether you are changing ECC memory settings for a physical GPU or a vGPU.

  • For a physical GPU, perform this task from the hypervisor host.
  • For a vGPU, perform this task from the VM to which the vGPU is assigned.
    Note: ECC memory must be enabled on the physical GPU on which the vGPUs reside.
Before you begin, ensure that NVIDIA Virtual GPU Manager is installed on your hypervisor. If you are changing ECC memory settings for a vGPU, also ensure that the NVIDIA AI Enterprise graphics driver is installed in the VM to which the vGPU is assigned.
  1. Use nvidia-smi to list the status of all physical GPUs or vGPUs, and check for ECC noted as enabled.
    # nvidia-smi -q
    
    ==============NVSMI LOG==============
    
    Timestamp                           : Mon Aug 21 18:36:45 2023
    Driver Version                      : 525.125.03
    
    Attached GPUs                       : 1
    GPU 0000:02:00.0
    
    [...]
    
        Ecc Mode
            Current                     : Enabled
            Pending                     : Enabled
    
    [...]
  2. Change the ECC status to off for each GPU for which ECC is enabled.
    • If you want to change the ECC status to off for all GPUs on your host machine or vGPUs assigned to the VM, run this command:
      # nvidia-smi -e 0
    • If you want to change the ECC status to off for a specific GPU or vGPU, run this command:
      # nvidia-smi -i id -e 0

      id is the index of the GPU or vGPU as reported by nvidia-smi.

      This example disables ECC for the GPU with index 0000:02:00.0.

      # nvidia-smi -i 0000:02:00.0 -e 0
  3. Reboot the host or restart the VM.
  4. Confirm that ECC is now disabled for the GPU or vGPU.
    # nvidia—smi —q
    
    ==============NVSMI LOG==============
    
    Timestamp                           : Mon Aug 21 18:37:53 2023
    Driver Version                      : 525.125.03
    
    Attached GPUs                       : 1
    GPU 0000:02:00.0
    [...]
    
        Ecc Mode
            Current                     : Disabled
            Pending                     : Disabled
    
    [...]
If you later need to enable ECC on your GPUs or vGPUs, follow the instructions in Enabling ECC Memory.

2.8.2. Enabling ECC Memory

If ECC memory is suitable for your workloads and is supported by your hypervisor software and GPUs, but is disabled on your GPUs or vGPUs, enable it.

Where to perform this task depends on whether you are changing ECC memory settings for a physical GPU or a vGPU.

  • For a physical GPU, perform this task from the hypervisor host.
  • For a vGPU, perform this task from the VM to which the vGPU is assigned.
    Note: ECC memory must be enabled on the physical GPU on which the vGPUs reside.
Before you begin, ensure that NVIDIA Virtual GPU Manager is installed on your hypervisor. If you are changing ECC memory settings for a vGPU, also ensure that the NVIDIA AI Enterprise graphics driver is installed in the VM to which the vGPU is assigned.
  1. Use nvidia-smi to list the status of all physical GPUs or vGPUs, and check for ECC noted as disabled.
    # nvidia-smi -q
    
    ==============NVSMI LOG==============
    
    Timestamp                           : Mon Aug 21 18:36:45 2023
    Driver Version                      : 525.125.03
    
    Attached GPUs                       : 1
    GPU 0000:02:00.0
    
    [...]
    
        Ecc Mode
            Current                     : Disabled
            Pending                     : Disabled
    
    [...]
  2. Change the ECC status to on for each GPU or vGPU for which ECC is enabled.
    • If you want to change the ECC status to on for all GPUs on your host machine or vGPUs assigned to the VM, run this command:
      # nvidia-smi -e 1
    • If you want to change the ECC status to on for a specific GPU or vGPU, run this command:
      # nvidia-smi -i id -e 1

      id is the index of the GPU or vGPU as reported by nvidia-smi.

      This example enables ECC for the GPU with index 0000:02:00.0.

      # nvidia-smi -i 0000:02:00.0 -e 1
  3. Reboot the host or restart the VM.
  4. Confirm that ECC is now enabled for the GPU or vGPU.
    # nvidia—smi —q
    
    ==============NVSMI LOG==============
    
    Timestamp                           : Mon Aug 21 18:37:53 2023
    Driver Version                      : 525.125.03
    
    Attached GPUs                       : 1
    GPU 0000:02:00.0
    [...]
    
        Ecc Mode
            Current                     : Enabled
            Pending                     : Enabled
    
    [...]
If you later need to disable ECC on your GPUs or vGPUs, follow the instructions in Disabling ECC Memory.

2.9. Configuring a vGPU VM for Use with NVIDIA GPUDirect Storage Technology

To use NVIDIA® GPUDirect Storage® technology with NVIDIA vGPU, you must install all the required software in the VM that is configured with NVIDIA vGPU.

Ensure that the prerequisites in Prerequisites for Using NVIDIA AI Enterprise are met.
  1. Install and configure the NVIDIA Virtual GPU Manager as explained in Installing and Configuring the NVIDIA Virtual GPU Manager for Red Hat Enterprise Linux KVM.
  2. As root, log in to the VM that you configured with NVIDIA vGPU in the previous step.
  3. Install the Mellanox OpenFabrics Enterprise Distribution for Linux (MLNX_OFED) in the VM as explained in Installation Procedure in Installing Mellanox OFED.

    In the command to run the installation script, specify the following options:

    • --with-nvmf
    • --with-nfsrdma
    • --enable-gds
    • --add-kernel-support
  4. Install the NVIDIA AI Enterprise graphics driver for Linux in the VM from a distribution-specific package.
    Note: GPUDirect Storage technology does not support installation of the NVIDIA AI Enterprise graphics driver for Linux from a .run file.

    Follow the instructions for the Linux distribution that is installed in the VM:

  5. Install NVIDIA CUDA Toolkit from a .run file, deselecting the CUDA driver when selecting the CUDA components to install.
    Note: To avoid overwriting the NVIDIA AI Enterprise graphics driver that you installed in the previous step, do not install NVIDIA CUDA Toolkit from a distribution-specific package.

    For instructions, refer to Runfile Installation in NVIDIA CUDA Installation Guide for Linux.

  6. Use the package manager of the Linux distribution that is installed in the VM to install the GPUDirect Storage technology packages, omitting the installation of the NVIDIA CUDA Toolkit packages.

    Follow the instructions in NVIDIA CUDA Installation Guide for Linux for the Linux distribution that is installed in the VM:

    • RHEL8/CentOS8

      In the step to install CUDA, execute only the command to include all GPUDirect Storage technology packages:
      sudo dnf install nvidia-gds
    • Ubuntu
      In the step to install CUDA, execute only the command to include all GPUDirect Storage technology packages:
      sudo apt-get install nvidia-gds

After you configure a vGPU VM for use with NVIDIA GPUDirect Storage technology, you can license the NVIDIA AI Enterprise licensed products that you are using. For instructions, refer to NVIDIA AI Enterprise Client Licensing User Guide.

3. Installing and Licensing NVIDIA AI Enterprise Software Components

3.1. Installing NVIDIA AI Enterprise Software Components by Using Kubernetes

Perform this task if you are using one of the following combinations of guest operating system and container platform:

  • Ubuntu with Kubernetes

Ensure that the following prerequisites are met:
  1. If you are using Kubernetes, ensure that:
    1. Kubernetes is installed in the VM.
    2. NVIDIA vGPU Manager is installed.
    3. NVIDIA vGPU License Server with licenses is installed.
  2. Helm is installed.
  3. You have generated your NGC API key for accessing the NVIDIA Enterprise Collection at the URL provided to you by NVIDIA.

3.1.2. Transforming Container Images for AI and Data Science Applications and Frameworks into Kubernetes Pods

The AI and data science applications and frameworks are distributed as NGC container images through the NGC private registry. If you are using Kubernetes or Red Hat OpenShift, you must transform each image that you want to use into a Kubernetes pod. Each container image contains the entire user-space software stack that is required to run the application or framework, namely, the CUDA libraries, cuDNN, any required Magnum IO components, TensorRT, and the framework.

3.2. Install NVIDIA AI Enterprise Software Components by Using Docker

Perform this task if you are using Ubuntu with Docker.

3.2.2. Installing NVIDIA AI Enterprise Software, Applications, and Deep Learning Framework Components by Using Docker

NVIDIA AI Enterprise software components in the infrastructure optimization and cloud native deployment layers are distributed through the NVIDIA AI Enterprise NGC private registry. Applications and deep learning framework components for NVIDIA AI Enterprise are distributed exclusively through the NGC Public Catalog.

The container image for each application or framework contains the entire user-space software stack that is required to run the application or framework, namely, the CUDA libraries, cuDNN, any required Magnum IO components, TensorRT, and the framework.

Ensure that you have completed the following tasks in NGC Private Registry User Guide:

Perform this task from the VM.
Obtain the Docker pull command for downloading each of the following applications and deep learning framework components from the listing for the application or component in the NGC Public Catalog.
  • Applications:
    • NVIDIA Clara Parabricks
    • NVIDIA DeepStream
    • MONAI - Medical Open Network for Artificial Intelligence
    • RAPIDS
    • RAPIDS Accelerator for Apache Spark
    • TAO
  • Deep learning framework components:
    • NVIDIA TensorRT
    • NVIDIA Triton Inference Server
    • PyTorch
    • TensorFlow 2
Obtain the command for downloading each of the following NVIDIA AI Enterprise software components from the listing for the component in the NVIDIA AI Enterprise NGC private registry.
  • GPU Operator
  • Network Operator
  • vGPU Guest Driver, Ubuntu 22.04

3.3. Installing and Licensing NVIDIA AI Enterprise Components Natively

3.3.1. Installing the NVIDIA AI Enterprise Graphics Driver on Windows

After you create a Windows VM on the hypervisor and boot the VM, the VM should boot to a standard Windows desktop in VGA mode at 800×600 resolution. You can use the Windows screen resolution control panel to increase the resolution to other standard resolutions, but to fully enable GPU operation, the NVIDIA AI Enterprise graphics driver must be installed. Windows guest VMs are supported on all NVIDIA vGPU types, namely: Q-series, C-series, B-series, and A-series NVIDIA vGPU types.

  1. Copy the NVIDIA Windows driver package to the guest VM where you are installing the driver.
  2. Execute the package to unpack and run the driver installer.
    Figure 13. NVIDIA driver installation

    Screen capture showing NVIDIA driver installation

  3. Click through the license agreement.
  4. Select Express Installation and click NEXT. After the driver installation is complete, the installer may prompt you to restart the platform.
  5. If prompted to restart the platform, do one of the following:
    • Select Restart Now to reboot the VM.
    • Exit the installer and reboot the VM when you are ready.
    After the VM restarts, it boots to a Windows desktop.
  6. Verify that the NVIDIA driver is running.
    1. Right-click on the desktop.
    2. From the menu that opens, choose NVIDIA Control Panel.
    3. In the NVIDIA Control Panel, from the Help menu, choose System Information.

      NVIDIA Control Panel reports the vGPU or physical GPU that is being used, its capabilities, and the NVIDIA driver version that is loaded.

      Figure 14. Verifying NVIDIA driver operation using NVIDIA Control Panel

      Screen capture showing the verification of NVIDIA driver operation using NVIDIA Control Panel

After you install the NVIDIA AI Enterprise graphics driver, you can license any NVIDIA AI Enterprise licensed products that you are using. For instructions, refer to NVIDIA AI Enterprise Client Licensing User Guide.

Note: The graphics driver for Windows in this release of NVIDIA AI Enterprise is distributed in a DCH-compliant package. A DCH-compliant package differs from a driver package that is not DCH compliant in the following ways:
  • The Windows registry key for license settings for a DCH-compliant package is different than the key for a driver package that is not DCH compliant. If you are upgrading from a driver package that is not DCH compliant in a VM that was previously licensed, you must reconfigure the license settings for the VM. Existing license settings are not propagated to the new Windows registry key for a DCH-compliant package.
  • NVIDIA System Management Interface, nvidia-smi, is installed in a folder that is in the default executable path.
  • The NVWMI binary files are installed in the Windows Driver Store under %SystemDrive%:\Windows\System32\DriverStore\FileRepository\.
  • NVWMI help information in Windows Help format is not installed with graphics driver for Windows guest OSes.

3.3.2. Installing the NVIDIA AI Enterprise Graphics Driver on Linux

The NVIDIA AI Enterprise graphics driver for Linux is distributed as a Debian package for Ubuntu distributions and as an RPM package for Red Hat distributions. The procedure for installing the driver is the same in a VM and on bare metal.

If you are using a Linux OS for which the Wayland display server protocol is enabled by default, disable it as explained in Disabling the Wayland Display Server Protocol for Red Hat Enterprise Linux.

How to install the NVIDIA AI Enterprisegraphics driver on Linux depends on the distribution format from which you are installing the driver. For detailed instructions, refer to:

3.3.2.1. Installing the NVIDIA AI Enterprise Graphics Driver on Ubuntu from a Debian Package

The NVIDIA AI Enterprise graphics driver for Ubuntu is distributed as a Debian package file.

This task requires sudo privileges.
  1. Copy the NVIDIA AI Enterprise Linux driver package, for example nvidia-linux-grid-525_525.125.06_amd64.deb, to the guest VM where you are installing the driver.
  2. Log in to the guest VM as a user with sudo privileges.
  3. Open a command shell and change to the directory that contains the NVIDIA AI Enterprise Linux driver package.
  4. From the command shell, run the command to install the package.
    $ sudo apt-get install ./nvidia-linux-grid-525_525.125.06_amd64.deb
  5. Verify that the NVIDIA driver is operational.
    1. Reboot the system and log in.
    2. After the system has rebooted, confirm that you can see your NVIDIA vGPU device in the output from the nvidia-smi command.
      $ nvidia-smi

3.3.2.2. Installing the NVIDIA AI Enterprise Graphics Driver on Red Hat Distributions from an RPM Package

The NVIDIA AI Enterprise graphics driver for Red Hat Distributions is distributed as an RPM package file.

This task requires root user privileges.
  1. Copy the NVIDIA AI Enterprise Linux driver package, for example nvidia-linux-grid-525_525.125.06_amd64.rpm, to the guest VM where you are installing the driver.
  2. Log in to the guest VM as a user with root user privileges.
  3. Open a command shell and change to the directory that contains the NVIDIA AI Enterprise Linux driver package.
  4. From the command shell, run the command to install the package.
    $ rpm -iv ./nvidia-linux-grid-525_525.125.06_amd64.rpm
  5. Verify that the NVIDIA driver is operational.
    1. Reboot the system and log in.
    2. After the system has rebooted, confirm that you can see your NVIDIA vGPU device in the output from the nvidia-smi command.
      $ nvidia-smi

3.3.2.3. Disabling the Nouveau Driver for NVIDIA Graphics Cards

If the Nouveau driver for NVIDIA graphics cards is present, disable it before installing the NVIDIA AI Enterprise graphics driver.
Note: If you are using SUSE Linux Enterprise Server, you can skip this task because the Nouveau driver is not present in SUSE Linux Enterprise Server.

Run the following command and if the command prints any output, the Nouveau driver is present and must be disabled.

$ lsmod | grep nouveau
  1. Create the file /etc/modprobe.d/blacklist-nouveau.conf with the following contents:
    blacklist nouveau
    options nouveau modeset=0
  2. Regenerate the kernel initial RAM file system (initramfs). The command to run to regenerate the kernel initramfs depends on the Linux distribution that you are using.
    Linux Distribution Command
    CentOS
    $ sudo dracut --force
    Debian
    $ sudo update-initramfs -u
    Red Hat Enterprise Linux
    $ sudo dracut --force
    Ubuntu
    $ sudo update-initramfs -u
  3. Reboot the host or guest VM.

3.3.2.4. Disabling the Wayland Display Server Protocol for Red Hat Enterprise Linux

Starting with Red Hat Enterprise Linux Desktop 8.0, the Wayland display server protocol is used by default on supported GPU and graphics driver configurations. However, the NVIDIA AI Enterprise graphics driver for Linux requires the X Window System. Before installing the driver, you must disable the Wayland display server protocol to revert to the X Window System.

Perform this task from the host or guest VM that is running Red Hat Enterprise Linux Desktop.
This task requires administrative access.
  1. In a plain text editor, edit the file /etc/gdm/custom.conf and remove the comment from the option WaylandEnable=false.
  2. Save your changes to /etc/gdm/custom.conf.
  3. Reboot the host or guest VM.

3.3.2.5. Disabling GSP Firmware

Some GPUs include a GPU System Processor (GSP), which may be used to offload GPU initialization and management tasks. In GPU pass through and bare-metal deployments on Linux, GSP is supported only for vCS. If you are using any other product in a GPU pass through or bare-metal deployment on Linux, you must disable the GSP firmware.
Note:

For NVIDIA vGPU deployments on Linux and all NVIDIA AI Enterprise deployments on Windows, omit this task.

GSP firmware is supported with NVIDIA vGPU deployments on GPUs that are based on the NVIDIA Ada Lovelace GPU architecture. For NVIDIA vGPU deployments on Linux and all NVIDIA AI Enterprise deployments on Windows on GPUs based on earlier GPU architectures, GSP is also not supported but GSP firmware is already disabled.

For each NVIDIA AI Enterprise product, the following table lists whether GSP is supported in deployments in which GSP firmware can be enabled. The table also summarizes the behavior of NVIDIA AI Enterprise if a VM or host requests a license when GSP firmware is enabled. The deployments in which GSP firmware can be enabled are GPU pass through and bare-metal deployments on Linux.

Product GSP License Request Error Message
vCS Supported Allowed Not applicable
vApps Not supported Blocked Printed
vWS Not supported Blocked Printed

When a license request is blocked, the following error message is written to the licensing event log file at the location given in NVIDIA AI Enterprise Client Licensing User Guide:

Invalid feature requested for the underlying GSP firmware configuration.
Disable GSP firmware to use this feature.

Perform this task on the VM to which the GPU is passed through or on the bare-metal host.

Ensure that the NVIDIA AI Enterprise graphics driver for Linux is installed on the VM or bare-metal host.
  1. Log in to the VM or bare-metal host and open a command shell.
  2. Determine whether GSP firmware is enabled.
    $ nvidia-smi -q
    • If GSP firmware is enabled, the command displays the GSP firmware version, for example:
      GSP Firmware Version                 : 525.125.06
    • Otherwise, the command displays N/A as the GSP firmware version.
  3. If GSP firmware is enabled, disable it by setting the NVIDIA module parameter NVreg_EnableGpuFirmware to 0.

    Set this parameter by adding the following entry to the /etc/modprobe.d/nvidia.conf file:

    options nvidia NVreg_EnableGpuFirmware=0

    If the /etc/modprobe.d/nvidia.conf file does not already exist, create it.

  4. Reboot the VM or bare-metal host.
If you later need to enable GSP firmware, set the NVIDIA module parameter NVreg_EnableGpuFirmware to 1.

3.3.3. Configuring a Licensed Client of NVIDIA License System

A client with a network connection obtains a license by leasing it from a NVIDIA License System service instance. The service instance serves the license to the client over the network from a pool of floating licenses obtained from the NVIDIA Licensing Portal. The license is returned to the service instance when the licensed client no longer requires the license.

Before configuring a licensed client, ensure that the following prerequisites are met:
  • The NVIDIA AI Enterprise graphics driver is installed on the client.
  • The client configuration token that you want to deploy on the client has been created from the NVIDIA Licensing Portal or the DLS as explained in NVIDIA License System User Guide.
  • Ports 443 and 80 in your firewall or proxy must be open to allow HTTPS traffic between a service instance and its the licensed clients. These ports must be open for both CLS instances and DLS instances.
    Note: For DLS releases before DLS 1.1, ports 8081 and 8082 were also required to be open to allow HTTPS traffic between a DLS instance and its licensed clients. Although these ports are no longer required, they remain supported for backward compatibility.

The graphics driver creates a default location in which to store the client configuration token on the client.

The process for configuring a licensed client is the same for CLS and DLS instances but depends on the OS that is running on the client.

3.3.3.1. Proxy Server Requirements and Firewall Rules for a CLS Instance

To enable communication between a licensed client and a CLS instance through a proxy server, the proxy server must meet certain requirements. To enable communication through a firewall, firewall rules that allow traffic on specific URLs through specific ports must be defined.

The processes for configuring a proxy server and defining firewall rules are separate from the process for configuring a CLS instance. Use the standard interfaces of the proxy server and the firewall that you are using to perform these processes.

Proxy Server Requirements for a CLS Instance
NVIDIA License System supports transparent proxy servers and non-transparent proxy servers.
  • A transparent proxy server identifies itself to the server and does not modify client requests and responses.
  • A non-transparent proxy server does not reveal the IP address of the client and modifies client requests and responses.

Any proxy server between a licensed client and a CLS instance must allow programmatic calls to the URL api.cls.licensing.nvidia.com.

Non-Transparent Proxy Server Support

NVIDIA License System supports both authenticated and unauthenticated non-transparent proxy servers.

The following authenticated proxy servers are supported:
  • Squid
The following authentication methods are supported for authenticated proxy servers:
  • Basic
  • Microsoft Windows Challenge/Response (Microsoft NTLM) (Windows clients only)
  • Kerberos (only for clients that are a member of an Active Directory domain)
Firewall Rules for a CLS Instance

To enable communication between a licensed client and a CLS instance through a firewall, firewall rules that allow traffic on the URLs through the ports specified in the following table must be defined.

URL Port Traffic
api.cls.licensing.nvidia.com 443
  • Licensing operations, namely, the borrowing, renewal, and return of a license.
  • Licensed client authentication
api.licensing.nvidia.com 80 License return from a Windows licensed client that has not been shut down cleanly

3.3.3.2. Configuring a Licensed Client on Windows with Default Settings

Perform this task from the client.

  1. Copy the client configuration token to the %SystemDrive%:\Program Files\NVIDIA Corporation\vGPU Licensing\ClientConfigToken folder.
  2. Restart the NvDisplayContainer service.

The NVIDIA service on the client should now automatically obtain a license from the CLS or DLS instance.

3.3.3.3. Configuring a Licensed Client on Linux with Default Settings

Perform this task from the client.

  1. As root, open the file /etc/nvidia/gridd.conf in a plain-text editor, such as vi.
    $ sudo vi /etc/nvidia/gridd.conf
    Note: You can create the /etc/nvidia/gridd.conf file by copying the supplied template file /etc/nvidia/gridd.conf.template.
  2. Add the FeatureType configuration parameter to the file /etc/nvidia/gridd.conf on a new line as FeatureType="value".

    value depends on the type of the GPU assigned to the licensed client that you are configuring.

    GPU Type Value
    NVIDIA vGPU 1. NVIDIA AI Enterprise automatically selects the correct type of license based on the vGPU type.
    Physical GPU The feature type of a GPU in pass-through mode or a bare-metal deployment:
    • 0: NVIDIA Virtual Applications
    • 2: NVIDIA RTX Virtual Workstation
    • 4: NVIDIA Virtual Compute Server

    This example shows how to configure a licensed Linux client for NVIDIA Virtual Compute Server.

    # /etc/nvidia/gridd.conf.template - Configuration file for NVIDIA Grid Daemon
    …
    # Description: Set Feature to be enabled
    # Data type: integer
    # Possible values:
    # 0 => for unlicensed state
    # 1 => for NVIDIA vGPU
    # 2 => for NVIDIA RTX Virtual Workstation
    # 4 => for NVIDIA Virtual Compute Server
    FeatureType=4
    ...
  3. Copy the client configuration token to the /etc/nvidia/ClientConfigToken directory.
  4. Ensure that the file access modes of the client configuration token allow the owner to read, write, and execute the token, and the group and others only to read the token.
    1. Determine the current file access modes of the client configuration token.
      # ls -l client-configuration-token-directory
    2. If necessary, change the mode of the client configuration token to 744.
      # chmod 744 client-configuration-token-directory/client_configuration_token_*.tok
    client-configuration-token-directory
    The directory to which you copied the client configuration token in the previous step.
  5. Save your changes to the /etc/nvidia/gridd.conf file and close the file.
  6. Restart the nvidia-gridd service.

The NVIDIA service on the client should now automatically obtain a license from the CLS or DLS instance.

3.3.3.4. Generating an Encrypted Credentials File

Some authentication methods require a licensed client to provide user credentials when the client authenticates with a proxy server. To enable the client to provide these credentials securely without input from a user, you must generate a file that contains these credentials in an encrypted form that the client can read.

The following authentication methods require an encrypted credentials file:
  • Basic authentication
  • Microsoft Windows Challenge/Response (NTLM) authentication for a client that is not a member of an Active Directory domain
How to generate an encrypted credentials file depends on the OS that client is running. For detailed instructions, refer to the following topics:

3.3.3.4.1. Generating an Encrypted Credentials File on Windows

Perform this task in a Windows PowerShell window as the Administrator user on the client.

  1. Change to the C:\Program Files\NVIDIA Corporation\vGPU Licensing\ProxySettings folder.
    PS C:\> cd "C:\Program Files\NVIDIA Corporation\vGPU Licensing\ProxySettings"
  2. Run the grid-proxy-credentials Windows PowerShell script.
    PS C:\> .\grid-proxy-credentials.ps1
  3. In the Select Output File Path window that opens, navigate to the directory in which you want to generate the credentials file, enter the file name, and click Save.



  4. When prompted in the Windows PowerShell window, specify the password of the user that will log in to the proxy server when the licensed client requests a license.
Provide the path to this file when configuring a licensed client that will use the file as explained in Configuring a Licensed Client on Windows with Default Settings.

3.3.3.4.2. Generating an Encrypted Credentials File on Linux

Perform this task in a Linux command shell on the client.

  1. Run the grid-proxy-credentials.sh command.
    # /usr/lib/nvidia/grid-proxy-credentials.sh -o output-file-path
    output-file-path
    The full path to the credentials file that you are generating. Ensure that the directory in the path exists.
    Tip: To get help information for this command, type /usr/lib/nvidia/grid-proxy-credentials.sh --help.
    This example creates the credentials file /etc/nvidia/proxy-credentials.dat.
    # /usr/lib/nvidia/grid-proxy-credentials.sh -o /etc/nvidia/proxy-credentials.dat
  2. When prompted, specify the password of the user that will log in to the proxy server when the licensed client requests a license.
Provide the path to this file when configuring a licensed client that will use the file as explained in Configuring a Licensed Client on Linux with Default Settings.

3.3.3.5. Verifying the NVIDIA AI Enterprise License Status of a Licensed Client

After configuring a client with an NVIDIA AI Enterprise license, verify the license status by displaying the licensed product name and status.

To verify the license status of a licensed client, run nvidia-smi with the –q or --query optionfrom the licensed client, not the hypervisor host. If the product is licensed, the expiration date is shown in the license status.

nvidia-smi -q
==============NVSMI LOG==============

Timestamp                                 : Wed Nov 23 10:52:59 2022
Driver Version                            : 525.60.06
CUDA Version                              : 12.0

Attached GPUs                             : 2
GPU 00000000:02:03.0
    Product Name                          : 
    Product Brand                         : NVIDIA Virtual Compute Server
    Product Architecture                  : Ampere
    Display Mode                          : Enabled
    Display Active                        : Disabled
    Persistence Mode                      : Enabled
    MIG Mode
        Current                           : Disabled
        Pending                           : Disabled
    Accounting Mode                       : Disabled
    Accounting Mode Buffer Size           : 4000
    Driver Model
        Current                           : N/A
        Pending                           : N/A
    Serial Number                         : N/A
    GPU UUID                              : GPU-ba5b1e9b-1dd3-11b2-be4f-98ef552f4216
    Minor Number                          : 0
    VBIOS Version                         : 00.00.00.00.00
    MultiGPU Board                        : No
    Board ID                              : 0x203
    Board Part Number                     : N/A
    GPU Part Number                       : 25B6-890-A1
    Module ID                             : N/A
    Inforom Version
        Image Version                     : N/A
        OEM Object                        : N/A
        ECC Object                        : N/A
        Power Management Object           : N/A
    GPU Operation Mode
        Current                           : N/A
        Pending                           : N/A
    GSP Firmware Version                  : N/A
    GPU Virtualization Mode
        Virtualization Mode               : VGPU
        Host VGPU Mode                    : N/A
    vGPU Software Licensed Product
        Product Name                      : NVIDIA Virtual Compute Server
        License Status                    : Licensed (Expiry: 2022-11-23 10:41:16 GMT)
    …
    …

3.3.4. Installing NVIDIA Container Toolkit

Use NVIDIA Container Toolkit to build and run GPU accelerated Docker containers. The toolkit includes a container runtime library and utilities to configure containers to use NVIDIA GPUs automatically.



Ensure that the following software is installed in the guest VM:

Note: You do not need to install NVIDIA CUDA Toolkit on the hypervisor host.
  1. Set up the GPG key and configure apt to use NVIDIA Container Toolkit packages in the file /etc/apt/sources.list.d/nvidia-docker.list.
    $ distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
    $ curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
    $ curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
  2. Download information from all configured sources about the latest versions of the packages and install the nvidia-container-toolkit package.
    $ sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit
  3. Restart the Docker service.
    $ sudo systemctl restart docker

3.3.5. Verifying the Installation of NVIDIA Container Toolkit

  1. Run the nvidia-smi command contained in the latest official NVIDIA CUDA Toolkit image.
    $ docker run --gpus all nvidia/cuda:11.0-base nvidia-smi
  2. Start a GPU-enabled container on any two available GPUs.
    $ docker run --gpus 2 nvidia/cuda:11.0-base nvidia-smi
  3. Start a GPU-enabled container on two specific GPUs identified by their index numbers.
    $ docker run --gpus '"device=1,2"' nvidia/cuda:10.0-base nvidia-smi
  4. Start a GPU-enabled container on two specific GPUs with one GPU identified by its UUID and the other GPU identified by its index number.
    $ docker run --gpus '"device=UUID-ABCDEF,1"' nvidia/cuda:11.0-base nvidia-smi
  5. Specify a GPU capability for the container.
    $ docker run --gpus all,capabilities=utility nvidia/cuda:11.0-base nvidia-smi

3.3.6. Installing Software Distributed as Container Images

The NGC container images accessed through the NVIDIA Enterprise Catalog include the AI and data science applications and frameworks. Each container image for an AI and data science application or framework contains the entire user-space software stack that is required to run the application or framework, namely, the CUDA libraries, cuDNN, any required Magnum IO components, TensorRT, and the framework.

Ensure that you have completed the following tasks in NGC Private Registry User Guide:

Perform this task from the VM.
Obtain the Docker pull command for downloading each of the following applications and deep learning framework components from the listing for the application or component in the NGC Public Catalog.
  • Applications:
    • NVIDIA Clara Parabricks
    • NVIDIA DeepStream
    • MONAI - Medical Open Network for Artificial Intelligence
    • RAPIDS
    • RAPIDS Accelerator for Apache Spark
    • TAO
  • Deep learning framework components:
    • NVIDIA TensorRT
    • NVIDIA Triton Inference Server
    • PyTorch
    • TensorFlow 2

3.3.7. Running ResNet-50 with TensorRT

  1. Launch the NVIDIA TensorRT container image on all GPUs in interactive mode, specifying that the container will be deleted when it is stopped.
    $ sudo docker run --gpus all -it --rm nvcr.io/nvaie/tensorrt:21.07-py3
  2. From within the container runtime, change to the directory that contains test data for the ResNet-50 convolutional neural network.
    # cd /workspace/tensorrt/data/resnet50
  3. Run the ResNet-50 convolutional neural network with FP32, FP16, and INT8 precision and confirm that each test is completed with the result PASSED.
    1. To run ResNet-50 with the default FP32 precision, run this command:
      # trtexec --duration=90 --workspace=1024 --percentile=99 --avgRuns=100 \
      --deploy=ResNet50_N2.prototxt --batch=1 --output=prob
    2. To run ResNet-50 with FP16 precision, add the --fp16 option:
      # trtexec --duration=90 --workspace=1024 --percentile=99 --avgRuns=100 \
      --deploy=ResNet50_N2.prototxt --batch=1 --output=prob --fp16
    3. To run ResNet-50 with INT8 precision, add the --int8 option:
      # trtexec --duration=90 --workspace=1024 --percentile=99 --avgRuns=100 \
      --deploy=ResNet50_N2.prototxt --batch=1 --output=prob --int8
  4. Press Ctrl+P, Ctrl+Q to exit the container runtime and return to the Linux command shell.

3.3.8. Running ResNet-50 with TensorFlow

  1. Launch the TensorFlow 1 container image on all GPUs in interactive mode, specifying that the container will be deleted when it is stopped.
    $ sudo docker run --gpus all -it --rm \
    nvcr.io/nvaie/tensorflow:21.07-tf1-py3
  2. From within the container runtime, change to the directory that contains test data for cnn example.
    # cd /workspace/nvidia-examples/cnn
  3. Run the ResNet-50 training test with FP16 precision.
    # python resnet.py --layers 50 -b 64 -i 200 -u batch --precision fp16
  4. Confirm that all operations on the application are performed correctly and that a set of results is reported when the test is completed.
  5. Press Ctrl+P, Ctrl+Q to exit the container runtime and return to the Linux command shell.

3.3.9. Optional: Updating NVIDIA Container Toolkit for a MIG-Enabled vGPU

To run containers on a MIG-enabled vGPU, you must update the NVIDIA Container Toolkit.
This task requires sudo privileges.
Perform this task from the guest VM in which you want to run containers on a MIG-enabled vGPU.
  1. Set up the GPG key and configure apt to use NVIDIA Container Toolkit packages in the file /etc/apt/sources.list.d/nvidia-docker.list.
    $ curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
        && distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
        && curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list \
        && sudo apt-get update
  2. Install the NVIDIA Container Toolkit packages and the packages on which it depends, and restart Docker.
    $ sudo apt-get install -y nvidia-docker2 \
        && sudo systemctl restart docker
  3. Test the installation of the NVIDIA Container Toolkit on the VM.
    sudo docker run –runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=mig-device nvidia/cuda nvidia-smi
    mig-device
    The ID of the MIG-enabled vGPU in one of the following formats:
    • MIG-gpu-uuid/gpu-instance-id/compute-instance-id
      gpu-uuid
      The UUID of the physical GPU, for example, GPU-786035d5-1e85-11b2-9fec-ac9c9a792daf.
      gpu-instance-id
      The index number the GPU instance on which the vGPU resides, for example, 0 for the first GPU instance.
      compute-instance-id
      The index number of the compute instance within the GPU instance, for example, 0 for the first compute instance.

      This example sets NVIDIA_VISIBLE_DEVICES for compute instance 0 on a MIG-enabled vGPU on GPU instance 0 of the physical GPU with UUID GPU-786035d5-1e85-11b2-9fec-ac9c9a792daf.

      NVIDIA_VISIBLE_DEVICES=MIG-GPU-786035d5-1e85-11b2-9fec-ac9c9a792daf/0/0
    • gpu-device-index:mig-device-index
      gpu-device-index
      The index number the physical GPU.
      mig-device-index
      The index number the GPU instance.

3.4. The Enterprise Catalog

The NVIDIA AI Enterprise Software Suite is distributed through the Enterprise Catalog. After you access the Enterprise Catalog, you will see the NVIDIA AI Enterprise Software Suite collection. Detailed documentation makes it easy to utilize the software, and if additional support is required, users can submit the ticket directly from the portal.

3.4.1. Accessing the NVIDIA AI Enterprise Collection

  1. Go to the NVIDIA AI Enterprise collection on NVIDIA NGC and, if prompted, log in.
  2. Under Enterprise Collections, scroll down to and click the NVIDIA AI Enterprise 3.2 collection.
  3. On the NVIDIA AI Enterprise 3.2 page that opens, click the Entities tab to review the software assets in the NVIDIA AI Enterprise 3.2 collection.
  4. Click the software asset that you are interested in to learn more about or download the asset.

3.4.2. Container Images

To pull AI and data science containers using Docker, follow these steps within the VM:
  1. Generate your API key.
  2. Access the Enterprise Catalog Container Registry.
    1. Log in to the NGC container registry.
      sudo docker login nvcr.io
    2. When prompted for your username, enter the text $oauthtoken.
      Username: $oauthtoken
    3. When prompted for your password, enter your NGC API key.
      Password: my-api-key
  3. For each AI or data science application that you are interested in, load the container.
    sudo docker pull nvcr.io/nvaie/tensorflow:21.02-tf2-py3

3.4.3. Helm Charts

The Enterprise Catalog contains Helm charts for use by NVIDIA GPU Operator and NVIDIA Network Operator to deploy and manage GPU resources and network resources for deployments based on Kubernetes.

  1. Go to the Enterprise Catalog.
  2. Click on the NVIDIA AI Enterprise Collection.
  3. Go to the Entities tab and select the Helm chart you are interested in.
  4. Here is how you download a Helm chart from the Enterprise Catalog.

3.4.4. Models

The Enterprise Catalog contains pretrained unencrypted models for MONAI and Tao Toolkit.

  1. Go to the Enterprise Catalog.
  2. Click on the NVIDIA AI Enterprise Collection.
  3. Go to the Entities tab and select the model you are interested in.
  4. On the page that opens for the model that you selected, click Download.

3.4.5. Resources

The NVIDIA AI Enterprise 3.2 collection contains the following resources:
  • GPU Operator
  • Network Operator
  • vGPU Guest Driver, Ubuntu 22.04
  1. Go to the Enterprise Catalog.
  2. Click the NVIDIA AI Enterprise 3.2 collection.
  3. Click the Entities tab and select the resource you are interested in.
  4. Download the resource directly from the UI or use the displayed wget or CLI commands.

3.4.6. Adding Additional Users from Your Organization to the Enterprise Catalog (Admins Only)

As an admin, you are responsible for giving members of your organization access to the Enterprise Catalog.

  1. Make sure you are signed in.
  2. Make sure to select your company's organization from the user menu on the top right.
  3. On the left side menu, select Organization and click on Users, then click the + icon at the bottom of the screen and then click the Invite New User icon.
  4. Provide the name and email address of the user you would like to add.
  5. Provision user roles for the new user:
    1. To give the new user access to the entities in the Enterprise Catalog, provide them with the user role NVIDIA AI Enterprise Viewer.
    2. To make the user an admin that can add additional users to the Enterprise Catalog, provision the user roles: NVIDIA AI Enterprise Viewer and User Admin.
    3. To give the user access to your organization’s Private Registry, see Accessing Your NGC Private Registry. Provisioning access to the Enterprise Catalog and your organization’s Private Registry can be done in one or two steps.

3.5. The NGC Private Registry

As an NVIDIA AI Enterprise user, you have exclusive access to your organization’s own NGC Private Registry, which gives authorized users within your organization privileges to store your company’s proprietary software and tools, including custom models, frameworks, and helm charts, in one location.

The complete NGC Private Registry user guide can be found here.

3.5.1. Accessing Your NGC Private Registry

  1. To access your NGC Private Registry, sign in with your NGC Account.
  2. In the top right corner, click your user account icon and select the orgname.



  3. To view artifacts in your NGC Private Registry, select Private Registry in the left-hand menu.



  4. You can access the content of the NGC Private Registry by selecting one of the entity types (Collections, Containers, Helm Charts, Models, Resources).
  5. To upload entities to your NGC Private Registry, click on Entity Creation Hub.

3.5.2. Managing Teams and Users

As an admin, you can add users to your organization’s NGC Private Registry and create teams within the NGC Private Registry.

Before adding users and teams, familiarize yourself with the following definitions of each role here.

3.5.2.1. Creating Teams

Creating teams allows users to share images within a team while keeping them invisible to other teams in the same organization. Only organization administrators can create teams.

Here is how you create a team.

3.5.2.2. Creating Users

As the organization administrator, you must create user accounts to allow others to use the NGC container registry within the organization.

Here is how you create a new user.

4. Configuring Multinode Scaling

Multinode scaling improves the performance of applications and frameworks, such as PyTorch and Tensorflow, that can run in a cluster of multiple hypervisor hosts.

Note:

Perform the tasks for configuring multinode scaling before performing the tasks in Getting Started with NVIDIA AI Enterprise.

The procedures for configuring switches and NICs apply to NVIDIA Mellanox NICs and switches. If you are using other makes of NICs and switches, consult the vendor's documentation for the products that you are using.

You are free to choose how to run your training jobs in a cluster. For information about the cluster architecture that can be used to run BERT training jobs, see Multi-node BERT User Guide.

4.1. Hardware and VM Configuration Requirements for Multinode Scaling

If you are configuring multinode scaling, your hardware and VM configuration must meet some specific requirements in addition to the requirements for a single node.

4.1.1. Hardware Requirements for Multinode Scaling

In addition to the requirements for a single node, the hardware used for multinode scaling must meet the following requirements:

  • An Ethernet NIC that supports RoCE is required in each VM used for multinode scaling. For best performance, NVIDIA recommends the NVIDIA® Mellanox® ConnectX®-6 Dx.

  • The hypervisor hosts must be connected to a network switch that supports RoCE. For best performance, NVIDIA recommends the NVIDIA Mellanox Spectrum switch.

  • One GPU is required for each VM.

    For best performance, NVIDIA recommends the NVIDIA A100 GPU.

  • Each GPU on each hypervisor host must be paired with a NIC in the same NUMA node.

4.2. Configuring NUMA Affinity for the VMs

To ensure that your multinode setup performs adequately, each GPU on each hypervisor host must be paired with a NIC in the same NUMA node. If a GPU is not paired with a NIC in the same NUMA node, reconfigure your server hardware to ensure that this prerequisite is met.

Examples of how to configure NUMA affinity for the VMs in a two-socket server are provided for the following configurations:.

  • Whole-server VM with two GPUs and two NICs across both NUMA nodes
  • Per-socket VM with one GPU and one NIC paired on a single NUMA node

The hardware configuration of the server is as follows:





4.2.1. Configuring NUMA Affinity for a Whole-Server VM with Two GPUs and Two NICs Across Both NUMA Nodes

The allocation of hardware resources to a VM that is assigned the whole server is as follows:

Perform this task on each hypervisor host.
  1. Determine the NUMA node to which the GPUs and NICs are attached.
    1. Determine the NUMA node to which the GPUs are attached.
      $ esxcli hardware pci list | grep -A 30 -B 10 NVIDIA
    2. Determine the NUMA node to which the NICs are attached.
      $ esxcli hardware pci list | grep -A 30 -B 10 Mellanox

    The following output describes two GPUs. One GPU is attached to NUMA node 0 and the other GPU is attached to NUMA node 1.

    #GPU 1
    0000:37:00.0
       Address: 0000:37:00.0
       Segment: 0x0000
       Bus: 0x37
       Slot: 0x00
       Function: 0x0
       VMkernel Name: vmgfx0
       Vendor Name: NVIDIA Corporation
       Device Name: NVIDIAA100-PCIE-40GB
       Configured Owner: VMkernel
       Current Owner: VMkernel
       Vendor ID: 0x10de
       Device ID: 0x20f1
       SubVendor ID: 0x10de
       SubDevice ID: 0x145f
       Device Class: 0x0302
       Device Class Name: 3D controller
       Programming Interface: 0x00
       Revision ID: 0xa1
       Interrupt Line: 0xff
       IRQ: 255
       Interrupt Vector: 0x00
       PCI Pin: 0x00
       Spawned Bus: 0x00
       Flags: 0x3001
       Module ID: 50
       Module Name: nvidia
       Chassis: 0
       Physical Slot: 2
       Slot Description: PCI-E Slot 2
       Device Layer Bus Address: s00000002.00
       Passthru Capable: true
       Parent Device: PCI 0:54:0:0
       Dependent Device: PCI 0:55:0:0
       Reset Method: Bridge reset
       FPT Sharable: true
       NUMA Node: 0
       Extended Device ID: 0
       Extended Device Name:
    
    #GPU 2
    0000:86:00.0
       Address: 0000:86:00.0
       Segment: 0x0000
       Bus: 0x86
       Slot: 0x00
       Function: 0x0
       VMkernel Name: vmgfx1
       Vendor Name: NVIDIA Corporation
       Device Name: NVIDIAA100-PCIE-40GB
       Configured Owner: VMkernel
       Current Owner: VMkernel
       Vendor ID: 0x10de
       Device ID: 0x20f1
       SubVendor ID: 0x10de
       SubDevice ID: 0x145f
       Device Class: 0x0302
       Device Class Name: 3D controller
       Programming Interface: 0x00
       Revision ID: 0xa1
       Interrupt Line: 0xff
       IRQ: 255
       Interrupt Vector: 0x00
       PCI Pin: 0x00
       Spawned Bus: 0x00
       Flags: 0x3001
       Module ID: 50
       Module Name: nvidia
       Chassis: 0
       Physical Slot: 5
       Slot Description: PCI-E Slot 5
       Device Layer Bus Address: s00000005.00
       Passthru Capable: true
       Parent Device: PCI 0:133:0:0
       Dependent Device: PCI 0:134:0:0
       Reset Method: Bridge reset
       FPT Sharable: true
       NUMA Node: 1
       Extended Device ID: 0
       Extended Device Name: 
  2. Set up vCPUs for the VM so that the VM has two sockets with the vCPU cores evenly divided between the sockets.
  3. With two GPUs and NICs in the VM across NUMA nodes, set the NUMA affinity in the VM configuration to include both NUMA nodes 0 and 1.
    numa.nodeAffinity = 0,1

4.2.2. Configuring NUMA Affinity for a Per-Socket VM with One GPU and One NIC on a Single NUMA Node

The allocation of hardware resources to the VMs that are each assigned one socket in a server is as follows:

Perform this task on each hypervisor host.
  1. Determine the NUMA node to which the GPUs and NICs are attached.
    1. Determine the NUMA node to which the GPUs are attached.
      $ esxcli hardware pci list | grep -A 30 -B 10 NVIDIA
    2. Determine the NUMA node to which the NICs are attached.
      $ esxcli hardware pci list | grep -A 30 -B 10 Mellanox

    The following output describes a GPU that is attached to NUMA node 0.

    0000:3b:02.3
       Address: 0000:3b:02.3
       Segment: 0x0000
       Bus: 0x3b
       Slot: 0x02
       Function: 0x3
       VMkernel Name: PF_0.59.0_VF_15
       Vendor Name: NVIDIA Corporation
       Device Name: NVIDIAA100-PCIE-40GB
       Configured Owner: 
       Current Owner: VMkernel
       Vendor ID: 0x10de
       Device ID: 0x20f1
       SubVendor ID: 0x10de
       SubDevice ID: 0x0000
       Device Class: 0x0302
       Device Class Name: 3D controller
       Programming Interface: 0x00
       Revision ID: 0xa1
       Interrupt Line: 0xff
       IRQ: 255
       Interrupt Vector: 0x00
    PCI Pin: 0xff
       Spawned Bus: 0x00
       Flags: 0x0001
       Module ID: 54
       Module Name: nvidia
       Chassis: 0
       Physical Slot: -1
       Slot Description:
       Device Layer Bus Address: s00000001.00.vf15
       Passthru Capable: true
       Parent Device: PCI 0:58:0:0
       Dependent Device: PCI 0:59:2:3
       Reset Method: Function reset
       FPT Sharable: true
       NUMA Node: 0
       Extended Device ID: 65535
       Extended Device Name: 
  2. For each GPU that you want to pair with a NIC, set the NUMA affinity in the VM configuration to the NUMA node to which the NIC and the GPU in the pair belong.
    numa.nodeAffinity = numa-node-value

4.3. Configuring RoCE on the NVIDIA Mellanox Spectrum Switch

The NVIDIA Mellanox Spectrum switch must be able to run RDMA over Converged Ethernet (RoCE) over a lossless network in DSCP-based QoS mode.

Perform this task from a host computer that has an Ethernet LAN connection to the switch.
  1. Use secure shell (SSH) to log in to the switch. To obtain the username and password for logging in to the switch, consult the documentation for the switch.
  2. Set the mode of the switch to RoCE.
    switch (config) # roce
  3. Create an isolated vLAN.
    switch (config) # interface vlan vlan-id

    The vLAN context is entered automatically after the vLAN is created.

    The following example creates a vLAN with the identifier 111.

    switch (config) # interface vlan 111
    switch (config vlan 111) # 
  4. Exit the vLAN context.
    switch (config vlan 111) # exit
  5. Place the NVIDIA ConnectX NICs into the created vLAN as access ports.
    switch (config) # interface ethernet port-range switchport access vlan-id

    This example puts four NVIDIA ConnectX NICs into the vLAN with the identifier 111 as access ports 1/1 - 1/4.

    switch (config) # interface ethernet 1/1-1/4 switchport access vlan 111
  6. Set the maximum transmission unit (MTU) frame size to 9216.
    1. Disable all the ports related to the interface.
      switch (config) # interface ethernet port-range shutdown
    2. Set the MTU frame size for the NVIDIA ConnectX NICs in the created vLAN to 9216.
      switch (config) # interface ethernet port-range mtu 9216
    3. Enable all the ports related to the interface.
      switch (config) # interface ethernet port-range no shutdown
  7. If your switch is running Cumulus Linux, enable RoCE with Cumulus Linux.

4.4. Enabling GPUDirect Technology for Peer-to-Peer Connections

Enabling GPUDirect® Technology for peer-to-peer connections involves enabling Address Translation Services (ATS) in the VMware ESXi VMkernel and modifying Access Control Services (ACS) settings for the VM.

Perform this task from each hypervisor host in your multinode cluster.
  1. As root, log in to the hypervisor host.
  2. Update the VMkernel settings.
    1. Enable Address Translation Services (ATS) in the boot options.
      [root@localhost:~] esxcli system settings kernel set -s atsSupport -v TRUE
    2. Reboot the hypervisor host.
    3. Confirm that ATS is enabled.
      [root@localhost:~] esxcli system settings kernel list -o atsSupport
      Name       Type Configured Runtime Default Description
      ---------- ---- ---------- ------- ------- -----------
      atsSupport Bool TRUE       TRUE    FALSE   Enable Support for PCIe          
                                                 ATS.
  3. Update the VM configuration.
    1. Set the option to enable peer-to-peer connections.
      pciPassthru.allowP2P=true
    2. Set the option to relax ACS settings for peer-to-peer connections.
      pciPassthru.RelaxACSforP2P=true

      When this option is set, VMware vSphere ESXi locates an ATS-capable pass-through device, finds its parent switch or root port, and enables the ACS Direct Translated bit.

4.5. Installing the Mellanox OFED Driver

Perform this task on each guest VM on each hypervisor host.
  1. Install the default version of python.
    $ sudo apt install python
  2. Download the compressed tar archive that contains the driver.
    $ wget \
    https://content.mellanox.com/ofed/MLNX_OFED-5.2-2.2.0.0/\
    MLNX_OFED_LINUX-5.2-2.2.0.0-ubuntu20.04-x86_64.tgz
    
  3. Extract the contents of the compressed tar archive that contains the driver.
    $ tar xvf MLNX_OFED_LINUX-5.2-2.2.0.0-ubuntu20.04-x86_64.tgz
  4. Change to the MLNX_OFED_LINUX-5.2-2.2.0.0-ubuntu20.04-x86_64 directory.
    $ cd MLNX_OFED_LINUX-5.2-2.2.0.0-ubuntu20.04-x86_64
  5. Run the script that installs the driver.
    $ sudo ./mlnxofedinstall

    During the installation process, OFED detects the ConnectX-6 NICs and updates the firmware.

  6. When the installation is complete, confirm that the versions of OFED are correct.
    1. Determine the OFED version.
      $ dpkg -l | grep mlnx-ofed
    2. Determine the firmware version.
      $ cat /sys/class/infiniband/mlx5*/fw_ver

    If the firmware is not updated, download the latest firmware, update the firmware manually, and install the Mellanox OFED driver again.

  7. Load the installed driver.
    $ sudo /etc/init.d/openibd restart

4.6. Enabling ATS on the NVIDIA ConnectX-6 DX NICs in a VM

Perform this task on each guest VM on each hypervisor host.
  1. Change the ATS configuration to enabled on each guest VM on the hypervisor host.
    1. Start Mellanox software tools.
      $ sudo mst start
    2. Determine whether ATS is enabled.
      $ sudo mlxconfig -d /dev/mst/mt4123_pciconf0 query | grep -i ATS

      If the installed version of the firmware supports ATS, output similar to the following example is displayed.

      ATS_ENABLED                         False(0)

      If no output is displayed, the installed version of the firmware does not support ATS. In this situation, update to a version of the firmware that supports ATS.

    3. If ATS is disabled, enable it.
      $sudo mlxconfig -d /dev/mst/mt4123_pciconf0 set ATS_ENABLED=true
      Device #1:
      ----------
      Device type:    ConnectX6
      Name:           MCX653105A-HDA_Ax
      Description:    ConnectX-6 VPI adapter card; HDR IB (200Gb/s) and 200GbE; single-port QSFP56; PCIe4.0 x16; tall bracket; ROHS R6
      Device:        /dev/mst/mt4123_pciconf0
       
      Configurations:           Next Boot   	New
      ATS_ENABLED               False(0)    	True(1)
      Apply new Configuration? (y/n) [n] : y
      Applying... Done!
      -I- Please reboot machine to load new configurations.
      
  2. After changing the ATS configuration to enabled on each guest VM on the node, turn off the power to the VMware vSphere ESXi host and turn the power back on again.
    Note:

    To apply the changed ATS configuration setting, you must turn off the power to the VMware vSphere ESXi host and turn the power back on again. Rebooting the host is insufficient to apply this change.

  3. Start VMware vCenter Server on the hypervisor host.
  4. For each VM on the node, perform the following steps:
    1. Turn on the power to the VM.
    2. Start Mellanox software tools.
      $ sudo mst start
    3. Determine whether ATS is enabled.
      $ sudo mlxconfig -d /dev/mst/mt4123_pciconf0 query | grep -i ATS

      If the installed version of the firmware supports ATS, output similar to the following example is displayed.

      ATS_ENABLED                         True(1)
    4. Obtain detailed information about all PCI buses and devices in the VM and confirm that the ATS capability of Mellanox ConnectX-6 device is shown as Enable+.
      $ sudo lspci -vvv
      ...
              Capabilities: [480 v1] Address Translation Service (ATS)
                      ATSCap: Invalidate Queue Depth: 00
                      ATSCtl: Enable+, Smallest Translation Unit: 00
      ...

4.7. Building and Installing the NVIDIA Peer Memory Driver

Perform this task on each guest VM on each hypervisor host.
  1. If necessary, install the latest stable upstream version of Git.
    1. Add the ppa:git-core/ppa repository to your list of package sources.
      $ sudo add-apt-repository ppa:git-core/ppa
    2. Download information from all configured sources about the latest versions of the packages.
      $ sudo apt update
    3. Install the git package.
      $ sudo apt install git
  2. Clone the Mellanox nv_peer_memory Git repository.
    $ git clone https://github.com/Mellanox/nv_peer_memory.git
  3. Change to the nv_peer_memory directory.
    $ cd nv_peer_memory/
  4. Build the NVIDIA peer memory driver software.
    $ ./build_module.sh
  5. Change to the /tmp directory.
    $ cd /tmp/
  6. Extract the NVIDIA peer memory driver software from the compressed tar archive that the build process created.
    $ tar xzf /tmp/nvidia-peer-memory_1.1.orig.tar.gz
  7. Change to the nvidia-peer-memory-1.1 directory.
    $ cd nvidia-peer-memory-1.1/
  8. Build the NVIDIA peer memory driver package.
    $ dpkg-buildpackage -us -uc
  9. Change to the parent of the current working directory.
    $ cd ..
  10. Install the driver package that you built.
    $ sudo dpkg -i nvidia-peer-memory_1.1-0_all.deb

5. Modifying a VM's NVIDIA vGPU Configuration

You can modify a VM's NVIDIA vGPU configuration by removing the NVIDIA vGPU configuration from a VM or by modifying GPU allocation policy.

5.1. Removing a VM’s NVIDIA vGPU Configuration

Remove a VM’s NVIDIA vGPU configuration when you no longer require the VM to use a virtual GPU.

5.1.1. Removing a vSphere VM’s vGPU Configuration

To remove a vSphere vGPU configuration from a VM:
  1. Select Edit settings after right-clicking on the VM in the vCenter Web UI.
  2. Select the Virtual Hardware tab.
  3. Mouse over the PCI Device entry showing NVIDIA GRID vGPU and click on the (X) icon to mark the device for removal.
  4. Click OK to remove the device and update the VM settings.

5.2. Modifying GPU Allocation Policy

VMware vSphere supports the breadth first and depth-first GPU allocation policies for vGPU-enabled VMs.

breadth-first
The breadth-first allocation policy attempts to minimize the number of vGPUs running on each physical GPU. Newly created vGPUs are placed on the physical GPU that can support the new vGPU and that has the fewest vGPUs already resident on it. This policy generally leads to higher performance because it attempts to minimize sharing of physical GPUs, but it may artificially limit the total number of vGPUs that can run.
depth-first
The depth-first allocation policy attempts to maximize the number of vGPUs running on each physical GPU. Newly created vGPUs are placed on the physical GPU that can support the new vGPU and that has the most vGPUs already resident on it. This policy generally leads to higher density of vGPUs, particularly when different types of vGPUs are being run, but may result in lower performance because it attempts to maximize sharing of physical GPUs.

By default, VMware vSphere ESXi uses the breadth-first allocation policy.

If the default GPU allocation policy does not meet your requirements for performance or density of vGPUs, you can change it.

5.2.1. Modifying GPU Allocation Policy on VMware vSphere

Before using the vSphere Web Client to change the allocation scheme, ensure that the ESXi host is running and that all VMs on the host are powered off.

  1. Log in to vCenter Server by using the vSphere Web Client.
  2. In the navigation tree, select your ESXi host and click the Configure tab.
  3. From the menu, choose Graphics and then click the Host Graphics tab.
  4. On the Host Graphics tab, click Edit.
    Figure 15. Breadth-first allocation scheme setting for vGPU-enabled VMs

    Screen capture of the Host Graphics tab in the VMware vCenter Web UI, showing the allocation scheme for vGPU-enabled VMs as breadth-first

  5. In the Edit Host Graphics Settings dialog box that opens, select these options and click OK.
    1. If not already selected, select Shared Direct.
    2. Select Group VMs on GPU until full.
    Figure 16. Host graphics settings for vGPU

    Screen capture showing the Edit Host Graphics Settings dialog box in the VMware vCenter Web UI for changing the allocation scheme for vGPU-enabled VMs

    After you click OK, the default graphics type changes to Shared Direct and the allocation scheme for vGPU-enabled VMs is breadth-first.

    Figure 17. Depth-first allocation scheme setting for vGPU-enabled VMs

    Screen capture of the Host Graphics tab in the VMware vCenter Web UI, showing the default graphics type as Shared Direct and the allocation scheme for vGPU-enabled VMs as depth-first

  6. Restart the ESXi host or the Xorg service on the host.

See also the following topics in the VMware vSphere documentation:

5.3. Migrating a VM Configured with vGPU

On some hypervisors, NVIDIA AI Enterprise supports migration of VMs that are configured with vGPU. Migration is supported for both time-sliced and MIG-backed vGPUs.

Before migrating a VM configured with vGPU, ensure that the following prerequisites are met:
  • The VM is configured with vGPU.
  • The VM is running.
  • The VM obtained a suitable vGPU license when it was booted.
  • The destination host has a physical GPU of the same type as the GPU where the vGPU currently resides.
  • If the physical GPU supports the Multi-Instance GPU (MIG) feature, the MIG configuration of the GPU on both the source and destination hosts must be identical.
  • ECC memory configuration (enabled or disabled) on both the source and destination hosts must be identical.
  • The GPU topologies (including NVLink widths) on both the source and destination hosts must be identical.
Note: vGPU migration is disabled for a VM for which any of the following NVIDIA CUDA Toolkit features is enabled:
  • Unified memory
  • Debuggers
  • Profilers
How to migrate a VM configured with vGPU depends on the hypervisor that you are using.

After migration, the vGPU type of the vGPU remains unchanged.

The time required for migration depends on the amount of frame buffer that the vGPU has. Migration for a vGPU with a large amount of frame buffer is slower than for a vGPU with a small amount of frame buffer.

5.3.1. Migrating a VM Configured with vGPU on VMware vSphere

NVIDIA AI Enterprise supports VMware vMotion for VMs that are configured with vGPU. VMware vMotion enables you to move a running virtual machine from one physical host machine to another host with very little disruption or downtime. For a VM that is configured with vGPU, the vGPU is migrated with the VM to an NVIDIA GPU on the other host. The NVIDIA GPUs on both host machines must be of the same type.

Perform this task in the VMware vSphere web client by using the Migration wizard.

Before migrating a VM configured with vGPU on VMware vSphere, ensure that the following prerequisites are met:
  1. Context-click the VM and from the menu that opens, choose Migrate.
  2. For the type of migration, select Change compute resource only and click Next. If you select Change both compute resource and storage, the time required for the migration increases.
  3. Select the destination host and click Next. The destination host must have a physical GPU of the same type as the GPU where the vGPU currently resides. Furthermore, the physical GPU must be capable of hosting the vGPU. If these requirements are not met, no available hosts are listed.
  4. Select the destination network and click Next.
  5. Select the migration priority level and click Next.
  6. Review your selections and click Finish.
For more information, see the following topics in the VMware documentation:

If NVIDIA vGPU migration is not configured, any attempt to migrate a VM with an NVIDIA vGPU fails and a window containing the following error message is displayed:

Compatibility Issue/Host
Migration was temporarily disabled due to another 
migration activity.
vGPU hot migration is not enabled.

The window appears as follows:



Screen capture showing the Compatibility Issue/Host error window

If you see this error, configure NVIDIA vGPU migration as explained in Configuring VMware vMotion with vGPU for VMware vSphere.

If your version of VMware vSpehere ESXi does not support vMotion for VMs configured with NVIDIA vGPU, any attempt to migrate a VM with an NVIDIA vGPU fails and a window containing the following error message is displayed:

Compatibility Issues
...
A required migration feature is not supported on the "Source" host 'host-name'.

A warning or error occurred when migrating the virtual machine.
Virtual machine relocation, or power on after relocation or cloning can fail if 
vGPU resources are not available on the destination host.

The window appears as follows:



Screen capture showing the Compatibility Issues error window

For details about which VMware vSphere versions, NVIDIA GPUs, and guest OS releases support suspend and resume, see NVIDIA AI Enterprise Release Notes.

5.3.2. Suspending and Resuming a VM Configured with vGPU on VMware vSphere

NVIDIA AI Enterprise supports suspend and resume for VMs that are configured with vGPU.

Perform this task in the VMware vSphere web client.

  • To suspend a VM, context-click the VM that you want to suspend, and from the context menu that pops up, choose Power > Suspend.
  • To resume a VM, context-click the VM that you want to resume, and from the context menu that pops up, choose Power > Power On.

5.4. Modifying a MIG-Backed vGPU's Configuration

If compute instances weren't created within the GPU instances when the GPU was configured for MIG-backed vGPUs, you can add the compute instances for an individual vGPU from within the guest VM. If you want to replace the compute instances that were created when the GPU was configured for MIG-backed vGPUs, you can delete them before adding the compute instances from within the guest VM.

Ensure that the following prerequisites are met:

  • You have root user privileges in the guest VM.
  • The GPU instance is not being used by any other processes, such as CUDA applications, monitoring applications, or the nvidia-smi command.

Perform this task in a guest VM command shell.

  1. Open a command shell as the root user in the guest VM. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. List the available GPU instance.
    $ nvidia-smi mig -lgi
    +----------------------------------------------------+
    | GPU instances:                                     |
    | GPU   Name          Profile  Instance   Placement  |
    |                       ID       ID       Start:Size |
    |====================================================|
    |   0  MIG 2g.10gb       0        0          0:8     |
    +----------------------------------------------------+
  3. Optional: If compute instances were created when the GPU was configured for MIG-backed vGPUs that you no longer require, delete them.
    $ nvidia-smi mig -dci -ci compute-instance-id -gi gpu-instance-id
    compute-instance-id
    The ID of the compute instance that you want to delete.
    gpu-instance-id
    The ID of the GPU instance from which you want to delete the compute instance.
    Note: If the GPU instance is being used by another process, this command fails. In this situation, stop all processes that are using the GPU instance and retry the command.

    This example deletes compute instance 0 from GPU instance 0 on GPU 0.

    $ nvidia-smi mig -dci -ci 0 -gi 0
    Successfully destroyed compute instance ID  0 from GPU  0 GPU instance ID  0
  4. List the compute instance profiles that are available for your GPU instance.
    $ nvidia-smi mig -lcip
    This example shows that one MIG 2g.10gb compute instance or two MIG 1c.2g.10gb compute instances can be created within the GPU instance.
    $ nvidia-smi mig -lcip
    +-------------------------------------------------------------------------------+
    | Compute instance profiles:                                                    |
    | GPU    GPU      Name          Profile  Instances   Exclusive      Shared      |
    |      Instance                   ID     Free/Total     SM      DEC   ENC   OFA |
    |        ID                                                     CE    JPEG      |
    |===============================================================================|
    |   0     0       MIG 1c.2g.10gb   0      2/2           14       1     0     0  |
    |                                                                2     0        |
    +-------------------------------------------------------------------------------+
    |   0     0       MIG 2g.10gb      1*     1/1           28       1     0     0  |
    |                                                                2     0        |
    +-------------------------------------------------------------------------------+
  5. Create the compute instances that you need within the available GPU instance.

    Create each compute instance individually by running the following command.

    $ nvidia-smi mig -cci compute-instance-profile-id -gi gpu-instance-id
    compute-instance-profile-id
    The compute instance profile ID that specifies the compute instance.
    gpu-instance-id
    The GPU instance ID that specifies the GPU instance within which you want to create the compute instance.
    Note: If the GPU instance is being used by another process, this command fails. In this situation, stop all processes that are using the GPU and retry the command.

    This example creates a MIG 2g.10gb compute instance on GPU instance 0.

    $ nvidia-smi mig -cci 1 -gi 0
    Successfully created compute instance ID  0 on GPU  0 GPU instance ID  0 using profile MIG 2g.10gb (ID  1)
    This example creates two MIG 1c.2g.10gb compute instances on GPU instance 0 by running the same command twice.
    $ nvidia-smi mig -cci 0 -gi 0
    Successfully created compute instance ID  0 on GPU  0 GPU instance ID  0 using profile MIG 1c.2g.10gb (ID  0)
    $ nvidia-smi mig -cci 0 -gi 0
    Successfully created compute instance ID  1 on GPU  0 GPU instance ID  0 using profile MIG 1c.2g.10gb (ID  0)
  6. Verify that the compute instances were created within the GPU instance. Use the nvidia-smi command for this purpose.

    This example confirms that a MIG 2g.10gb compute instance was created on GPU instance 0.

    nvidia-smi
    Mon Aug 21 19:01:24 2023
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.03   Driver Version: 525.125.03  CUDA Version:  12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  GRID A100X-2-10C     On  | 00000000:00:08.0 Off |                   On |
    | N/A   N/A    P0    N/A /  N/A |   1058MiB / 10235MiB |     N/A      Default |
    |                               |                      |              Enabled |
    +-------------------------------+----------------------+----------------------+
    
    +-----------------------------------------------------------------------------+
    | MIG devices:                                                                |
    +------------------+----------------------+-----------+-----------------------+
    | GPU  GI  CI  MIG |         Memory-Usage |        Vol|         Shared        |
    |      ID  ID  Dev |           BAR1-Usage | SM     Unc| CE  ENC  DEC  OFA  JPG|
    |                  |                      |        ECC|                       |
    |==================+======================+===========+=======================|
    |  0    0   0   0  |   1058MiB / 10235MiB | 28      0 |  2   0    1    0    0 |
    |                  |      0MiB /  4096MiB |           |                       |
    +------------------+----------------------+-----------+-----------------------+
    
    +-----------------------------------------------------------------------------+
    | Processes:                                                                  |
    |  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
    |        ID   ID                                                   Usage      |
    |=============================================================================|
    |  No running processes found                                                 |
    +-----------------------------------------------------------------------------+

    This example confirms that two MIG 1c.2g.10gb compute instances were created on GPU instance 0.

    $ nvidia-smi
    Mon Aug 21 19:01:24 2023
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 525.125.03   Driver Version: 525.125.03  CUDA Version:  12.0     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |                               |                      |               MIG M. |
    |===============================+======================+======================|
    |   0  GRID A100X-2-10C     On  | 00000000:00:08.0 Off |                   On |
    | N/A   N/A    P0    N/A /  N/A |   1058MiB / 10235MiB |     N/A      Default |
    |                               |                      |              Enabled |
    +-------------------------------+----------------------+----------------------+
    
    +-----------------------------------------------------------------------------+
    | MIG devices:                                                                |
    +------------------+----------------------+-----------+-----------------------+
    | GPU  GI  CI  MIG |         Memory-Usage |        Vol|         Shared        |
    |      ID  ID  Dev |           BAR1-Usage | SM     Unc| CE  ENC  DEC  OFA  JPG|
    |                  |                      |        ECC|                       |
    |==================+======================+===========+=======================|
    |  0    0   0   0  |   1058MiB / 10235MiB | 14      0 |  2   0    1    0    0 |
    |                  |      0MiB /  4096MiB |           |                       |
    +------------------+                      +-----------+-----------------------+
    |  0    0   1   1  |                      | 14      0 |  2   0    1    0    0 |
    |                  |                      |           |                       |
    +------------------+----------------------+-----------+-----------------------+
    
    +-----------------------------------------------------------------------------+
    | Processes:                                                                  |
    |  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
    |        ID   ID                                                   Usage      |
    |=============================================================================|
    |  No running processes found                                                 |
    +-----------------------------------------------------------------------------+
    

5.5. Enabling Unified Memory for a vGPU

Unified memory is disabled by default. If used, you must enable unified memory individually for each vGPU that requires it by setting a vGPU plugin parameter. How to enable unified memory for a vGPU depends on the hypervisor that you are using.

5.5.1. Enabling Unified Memory for a vGPU on Red Hat Enterprise Linux KVM

On Red Hat Enterprise Linux KVM, enable unified memory by setting the enable_uvm vGPU plugin parameter.

Ensure that the mdev device file that represents the vGPU has been created as explained in Creating an NVIDIA vGPU on a Linux with KVM Hypervisor.
Perform this task for each vGPU that requires unified memory.
Set the enable_uvm vGPU plugin parameter for the mdev device file that represents the vGPU to 1 as explained in Setting vGPU Plugin Parameters on a Linux with KVM Hypervisor.

5.5.2. Enabling Unified Memory for a vGPU on VMware vSphere

On VMware vSphere, enable unified memory by setting the pciPassthruvgpu-id.cfg.enable_uvm configuration parameter in advanced VM attributes.

Ensure that the VM to which the vGPU is assigned is powered off.
Perform this task in the vSphere Client for each vGPU that requires unified memory.

In advanced VM attributes, set the pciPassthruvgpu-id.cfg.enable_uvm vGPU plugin parameter for the vGPU to 1 as explained in Setting vGPU Plugin Parameters on VMware vSphere.

vgpu-id
A positive integer that identifies the vGPU assigned to a VM. For the first vGPU assigned to a VM, vgpu-id is 0. For example, if two vGPUs are assigned to a VM and you are enabling unified memory for both vGPUs, set pciPassthru0.cfg.enable_uvm and pciPassthru1.cfg.enable_uvm to 1.

5.6. Enabling NVIDIA CUDA Toolkit Development Tools for NVIDIA vGPU

By default, NVIDIA CUDA Toolkit development tools are disabled on NVIDIA vGPU. If used, you must enable NVIDIA CUDA Toolkit development tools individually for each VM that requires them by setting vGPU plugin parameters. One parameter must be set for enabling NVIDIA CUDA Toolkit debuggers and a different parameter must be set for enabling NVIDIA CUDA Toolkit profilers.

5.6.1. Enabling NVIDIA CUDA Toolkit Debuggers for NVIDIA vGPU

By default, NVIDIA CUDA Toolkit debuggers are disabled. If used, you must enable them for each vGPU VM that requires them by setting a vGPU plugin parameter. How to set the parameter to enable NVIDIA CUDA Toolkit debuggers for a vGPU VM depends on the hypervisor that you are using.

You can enable NVIDIA CUDA Toolkit debuggers for any number of VMs configured with vGPUs on the same GPU. When NVIDIA CUDA Toolkit debuggers are enabled for a VM, the VM cannot be migrated.

Perform this task for each VM for which you want to enable NVIDIA CUDA Toolkit debuggers.

Enabling NVIDIA CUDA Toolkit Debuggers for NVIDIA vGPU on Red Hat Enterprise Linux KVM

Set the enable_debugging vGPU plugin parameter for the mdev device file that represents the vGPU that is assigned to the VM to 1 as explained in Setting vGPU Plugin Parameters on a Linux with KVM Hypervisor.

The setting of this parameter is preserved after a guest VM is restarted. However, this parameter is reset to its default value after the hypervisor host is restarted.

Enabling NVIDIA CUDA Toolkit Debuggers for NVIDIA vGPU on on VMware vSphere

Ensure that the VM for which you want to enable NVIDIA CUDA Toolkit debuggers is powered off.

In advanced VM attributes, set the pciPassthruvgpu-id.cfg.enable_debugging vGPU plugin parameter for the vGPU that is assigned to the VM to 1 as explained in Setting vGPU Plugin Parameters on VMware vSphere.

vgpu-id
A positive integer that identifies the vGPU assigned to the VM. For the first vGPU assigned to a VM, vgpu-id is 0. For example, if two vGPUs are assigned to a VM and you are enabling debuggers for both vGPUs, set pciPassthru0.cfg.enable_debugging and pciPassthru1.cfg.enable_debugging to 1.

The setting of this parameter is preserved after a guest VM is restarted. However, this parameter is reset to its default value after the hypervisor host is restarted.

5.6.2. Enabling NVIDIA CUDA Toolkit Profilers for NVIDIA vGPU

By default, only GPU workload trace is enabled. If you want to use all NVIDIA CUDA Toolkit profiler features that NVIDIA vGPU supports, you must enable them for each vGPU VM that requires them.

Note: Enabling profiling for a VM gives the VM access to the GPU’s global performance counters, which may include activity from other VMs executing on the same GPU. Enabling profiling for a VM also allows the VM to lock clocks on the GPU, which impacts all other VMs executing on the same GPU.

5.6.2.1. Supported NVIDIA CUDA Toolkit Profiler Features

You can enable the following NVIDIA CUDA Toolkit profiler features for a vGPU VM:

  • NVIDIA Nsight™ Compute
  • NVIDIA Nsight Systems
  • CUDA Profiling Tools Interface (CUPTI)

5.6.2.2. Clock Management for a vGPU VM for Which NVIDIA CUDA Toolkit Profilers Are Enabled

Clocks are not locked for periodic sampling use cases such as NVIDIA Nsight Systems profiling.

Clocks are locked for multipass profiling such as:

  • NVIDIA Nsight Compute kernel profiling
  • CUPTI range profiling

Clocks are locked automatically when profiling starts and are unlocked automatically when profiling ends.

5.6.2.3. Limitations on the Use of NVIDIA CUDA Toolkit Profilers with NVIDIA vGPU

The following limitations apply when NVIDIA CUDA Toolkit profilers are enabled for NVIDIA vGPU:

  • NVIDIA CUDA Toolkit profilers can be used on only one VM at a time.
  • Multiple CUDA contexts cannot be profiled simultaneously.
  • Profiling data is collected separately for each context.
  • A VM for which NVIDIA CUDA Toolkit profilers are enabled cannot be migrated.

Because NVIDIA CUDA Toolkit profilers can be used on only one VM at a time, you should enable them for only one VM assigned a vGPU on a GPU. However, NVIDIA AI Enterprise cannot enforce this requirement. If NVIDIA CUDA Toolkit profilers are enabled on more than one VM assigned a vGPU on a GPU, profiling data is collected only for the first VM to start the profiler.

5.6.2.4. Enabling NVIDIA CUDA Toolkit Profilers for a vGPU VM

You enable NVIDIA CUDA Toolkit profilers for a vGPU VM by setting a vGPU plugin parameter. How to set the parameter to enable NVIDIA CUDA Toolkit profilers for a vGPU VM depends on the hypervisor that you are using.

Perform this task for the VM for which you want to enable NVIDIA CUDA Toolkit profilers.

Enabling NVIDIA CUDA Toolkit Profilers for NVIDIA vGPU on Red Hat Enterprise Linux KVM

Set the enable_profiling vGPU plugin parameter for the mdev device file that represents the vGPU that is assigned to the VM to 1 as explained in Setting vGPU Plugin Parameters on a Linux with KVM Hypervisor.

The setting of this parameter is preserved after a guest VM is restarted. However, this parameter is reset to its default value after the hypervisor host is restarted.

Enabling NVIDIA CUDA Toolkit Profilers for NVIDIA vGPU on on VMware vSphere

Ensure that the VM for which you want to enable NVIDIA CUDA Toolkit profilers is powered off.

In advanced VM attributes, set the pciPassthruvgpu-id.cfg.enable_profiling vGPU plugin parameter for the vGPU that is assigned to the VM to 1 as explained in Setting vGPU Plugin Parameters on VMware vSphere.

vgpu-id
A positive integer that identifies the vGPU assigned to the VM. For the first vGPU assigned to a VM, vgpu-id is 0. For example, if two vGPUs are assigned to a VM and you are enabling profilers for the second vGPU, set pciPassthru1.cfg.enable_profiling to 1.

The setting of this parameter is preserved after a guest VM is restarted. However, this parameter is reset to its default value after the hypervisor host is restarted.

5.7. Enabling the TCC Driver Model for a vGPU

The Tesla Compute Cluster (TCC) driver model supports CUDA C/C++ applications. This model is optimized for compute applications and reduces kernel launch times on Windows. By default, the driver model of a vGPU that is assigned to a Windows VM is Windows Display Driver Model (WDDM). If you want to use the TCC driver model, you must enable it explicitly.

This task requires administrator privileges.
Perform this task from the VM to which the vGPU is assigned.
Note: Only Q-series vGPUs support the TCC driver model.
  1. Log on to the VM to which the vGPU is assigned.
  2. Set the driver model of the vGPU to the TCC driver model.
    nvidia-smi -g vgpu-id -dm 1
    vgpu-id
    The ID of the vGPU for which you want to enable the TCC driver model. If the -g is omitted, the TCC driver model is enabled for all vGPUs that are assigned to the VM.
  3. Reboot the VM.

6. Monitoring GPU Performance

NVIDIA AI Enterprise enables you to monitor the performance of physical GPUs and virtual GPUs from the hypervisor and from within individual guest VMs.

6.1. NVIDIA System Management Interface nvidia-smi

NVIDIA System Management Interface, nvidia-smi, is a command-line tool that reports management information for NVIDIA GPUs.

The nvidia-smi tool is included in the following packages:

  • NVIDIA Virtual GPU Manager package for each supported hypervisor
  • NVIDIA driver package for each supported guest OS

The scope of the reported management information depends on where you run nvidia-smi from:

  • From a hypervisor command shell, such as the VMware ESXi host shell, nvidia-smi reports management information for NVIDIA physical GPUs and virtual GPUs present in the system.

  • From a guest VM, nvidia-smi retrieves usage statistics for vGPUs or pass-through GPUs that are assigned to the VM.

6.2. Using nvidia-smi to Monitor GPU Performance from a Hypervisor

You can get management information for the NVIDIA physical GPUs and virtual GPUs present in the system by running nvidia-smi from a hypervisor command shell such as the Citrix Hypervisor dom0 shell or the VMware ESXi host shell.

Without a subcommand, nvidia-smi provides management information for physical GPUs. To examine virtual GPUs in more detail, use nvidia-smi with the vgpu subcommand.

From the command line, you can get help information about the nvidia-smi tool and the vgpu subcommand.

Help Information Command
A list of subcommands supported by the nvidia-smi tool. Note that not all subcommands apply to GPUs that support NVIDIA AI Enterprise. nvidia-smi -h
A list of all options supported by the vgpu subcommand. nvidia-smi vgpu –h

6.2.1. Getting a Summary of all Physical GPUs in the System

To get a summary of all physical GPUs in the system, along with PCI bus IDs, power state, temperature, current memory usage, and so on, run nvidia-smi without additional arguments.

Each vGPU instance is reported in the Compute processes section, together with its physical GPU index and the amount of frame-buffer memory assigned to it.

In the example that follows, three vGPUs are running in the system: One vGPU is running on each of the physical GPUs 0, 1, and 2.

[root@vgpu ~]# nvidia-smi
Fri Aug 17 09:26:18 2024
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.125.03               Driver Version: 525.125.03              |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M60           On   | 0000:83:00.0     Off |                  Off |
| N/A   31C    P8    23W / 150W |   1889MiB /  8191MiB |      7%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla M60           On   | 0000:84:00.0     Off |                  Off |
| N/A   26C    P8    23W / 150W |    926MiB /  8191MiB |      9%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla M10           On   | 0000:8A:00.0     Off |                  N/A |
| N/A   23C    P8    10W /  53W |   1882MiB /  8191MiB |     12%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Tesla M10           On   | 0000:8B:00.0     Off |                  N/A |
| N/A   26C    P8    10W /  53W |     10MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   4  Tesla M10           On   | 0000:8C:00.0     Off |                  N/A |
| N/A   34C    P8    10W /  53W |     10MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   5  Tesla M10           On   | 0000:8D:00.0     Off |                  N/A |
| N/A   32C    P8    10W /  53W |     10MiB /  8191MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|    0     11924  C+G   /usr/lib64/xen/bin/vgpu                       1856MiB |
|    1     11903  C+G   /usr/lib64/xen/bin/vgpu                        896MiB |
|    2     11908  C+G   /usr/lib64/xen/bin/vgpu                       1856MiB |
+-----------------------------------------------------------------------------+
[root@vgpu ~]#

6.2.2. Getting a Summary of all vGPUs in the System

To get a summary of the vGPUs currently that are currently running on each physical GPU in the system, run nvidia-smi vgpu without additional arguments.

[root@vgpu ~]# nvidia-smi vgpu
Fri Aug 17 09:27:06 2024
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.125.03               Driver Version: 525.125.03              |
|-------------------------------+--------------------------------+------------+
| GPU  Name                     | Bus-Id                         | GPU-Util   |
|      vGPU ID    Name          | VM ID    VM Name               | vGPU-Util  |
|===============================+================================+============|
|   0  Tesla M60                | 0000:83:00.0                   |   7%       |
|      11924      GRID M60-2Q   | 3        Win7-64 GRID test 2   |       6%   |
+-------------------------------+--------------------------------+------------+
|   1  Tesla M60                | 0000:84:00.0                   |   9%       |
|      11903      GRID M60-1B   | 1        Win8.1-64 GRID test 3 |       8%   |
+-------------------------------+--------------------------------+------------+
|   2  Tesla M10                | 0000:8A:00.0                   |  12%       |
|      11908      GRID M10-2Q   | 2        Win7-64 GRID test 1   |      10%   |
+-------------------------------+--------------------------------+------------+
|   3  Tesla M10                | 0000:8B:00.0                   |   0%       |
+-------------------------------+--------------------------------+------------+
|   4  Tesla M10                | 0000:8C:00.0                   |   0%       |
+-------------------------------+--------------------------------+------------+
|   5  Tesla M10                | 0000:8D:00.0                   |   0%       |
+-------------------------------+--------------------------------+------------+
[root@vgpu ~]#

6.2.3. Getting Physical GPU Details

To get detailed information about all the physical GPUs on the platform, run nvidia-smi with the –q or --query option.

[root@vgpu ~]# nvidia-smi -q
==============NVSMI LOG==============

Timestamp                                 : Tue Nov 22 10:33:26 2022
Driver Version                            : 525.60.06
CUDA Version                              : Not Found
vGPU Driver Capability
        Heterogenous Multi-vGPU           : Supported

Attached GPUs                             : 3
GPU 00000000:C1:00.0
    Product Name                          : Tesla T4
    Product Brand                         : NVIDIA
    Product Architecture                  : Turing
    Display Mode                          : Enabled
    Display Active                        : Disabled
    Persistence Mode                      : Enabled
    vGPU Device Capability
        Fractional Multi-vGPU             : Supported
        Heterogeneous Time-Slice Profiles : Supported
        Heterogeneous Time-Slice Sizes    : Not Supported
    MIG Mode
        Current                           : N/A
        Pending                           : N/A
    Accounting Mode                       : Enabled
    Accounting Mode Buffer Size           : 4000
    Driver Model
        Current                           : N/A
        Pending                           : N/A
    Serial Number                         : 1321120031291
    GPU UUID                              : GPU-9084c1b2-624f-2267-4b66-345583fbd981
    Minor Number                          : 1
    VBIOS Version                         : 90.04.38.00.03
    MultiGPU Board                        : No
    Board ID                              : 0xc100
    Board Part Number                     : 900-2G183-0000-001
    GPU Part Number                       : 1EB8-895-A1
    Module ID                             : 0
    Inforom Version
        Image Version                     : G183.0200.00.02
        OEM Object                        : 1.1
        ECC Object                        : 5.0
        Power Management Object           : N/A
    GPU Operation Mode
        Current                           : N/A
        Pending                           : N/A
    GSP Firmware Version                  : N/A
    GPU Virtualization Mode
        Virtualization Mode               : Host VGPU
        Host VGPU Mode                    : Non SR-IOV
    IBMNPU
        Relaxed Ordering Mode             : N/A
    PCI
        Bus                               : 0xC1
        Device                            : 0x00
        Domain                            : 0x0000
        Device Id                         : 0x1EB810DE
        Bus Id                            : 00000000:C1:00.0
        Sub System Id                     : 0x12A210DE
        GPU Link Info
            PCIe Generation
                Max                       : 3
                Current                   : 1
                Device Current            : 1
                Device Max                : 3
                Host Max                  : N/A
            Link Width
                Max                       : 16x
                Current                   : 16x
        Bridge Chip
            Type                          : N/A
            Firmware                      : N/A
        Replays Since Reset               : 0
        Replay Number Rollovers           : 0
        Tx Throughput                     : 0 KB/s
        Rx Throughput                     : 0 KB/s
        Atomic Caps Inbound               : N/A
        Atomic Caps Outbound              : N/A
    Fan Speed                             : N/A
    Performance State                     : P8
    Clocks Throttle Reasons
        Idle                              : Active
        Applications Clocks Setting       : Not Active
        SW Power Cap                      : Not Active
        HW Slowdown                       : Not Active
            HW Thermal Slowdown           : Not Active
            HW Power Brake Slowdown       : Not Active
        Sync Boost                        : Not Active
        SW Thermal Slowdown               : Not Active
        Display Clock Setting             : Not Active
    FB Memory Usage
        Total                             : 15360 MiB
        Reserved                          : 0 MiB
        Used                              : 3859 MiB
        Free                              : 11500 MiB
    BAR1 Memory Usage
        Total                             : 256 MiB
        Used                              : 17 MiB
        Free                              : 239 MiB
    Compute Mode                          : Default
    Utilization
        Gpu                               : 0 %
        Memory                            : 0 %
        Encoder                           : 0 %
        Decoder                           : 0 %
    Encoder Stats
        Active Sessions                   : 0
        Average FPS                       : 0
        Average Latency                   : 0
    FBC Stats
        Active Sessions                   : 0
        Average FPS                       : 0
        Average Latency                   : 0
    Ecc Mode
        Current                           : Enabled
        Pending                           : Enabled
    ECC Errors
        Volatile
            SRAM Correctable              : 0
            SRAM Uncorrectable            : 0
            DRAM Correctable              : 0
            DRAM Uncorrectable            : 0
        Aggregate
            SRAM Correctable              : 0
            SRAM Uncorrectable            : 0
            DRAM Correctable              : 0
            DRAM Uncorrectable            : 0
    Retired Pages
        Single Bit ECC                    : 0
        Double Bit ECC                    : 0
        Pending Page Blacklist            : No
    Remapped Rows                         : N/A
    Temperature
        GPU Current Temp                  : 35 C
        GPU Shutdown Temp                 : 96 C
        GPU Slowdown Temp                 : 93 C
        GPU Max Operating Temp            : 85 C
        GPU Target Temperature            : N/A
        Memory Current Temp               : N/A
        Memory Max Operating Temp         : N/A
    Power Readings
        Power Management                  : Supported
        Power Draw                        : 16.57 W
        Power Limit                       : 70.00 W
        Default Power Limit               : 70.00 W
        Enforced Power Limit              : 70.00 W
        Min Power Limit                   : 60.00 W
        Max Power Limit                   : 70.00 W
    Clocks
        Graphics                          : 300 MHz
        SM                                : 300 MHz
        Memory                            : 405 MHz
        Video                             : 540 MHz
    Applications Clocks
        Graphics                          : 585 MHz
        Memory                            : 5001 MHz
    Default Applications Clocks
        Graphics                          : 585 MHz
        Memory                            : 5001 MHz
    Deferred Clocks
        Memory                            : N/A
    Max Clocks
        Graphics                          : 1590 MHz
        SM                                : 1590 MHz
        Memory                            : 5001 MHz
        Video                             : 1470 MHz
    Max Customer Boost Clocks
        Graphics                          : 1590 MHz
    Clock Policy
        Auto Boost                        : N/A
        Auto Boost Default                : N/A
    Voltage
        Graphics                          : N/A
    Fabric
        State                             : N/A
        Status                            : N/A
    Processes
        GPU instance ID                   : N/A
        Compute instance ID               : N/A
        Process ID                        : 2103065
            Type                          : C+G
            Name                          : Win11SV2_View87                   
            Used GPU Memory               : 3810 MiB
[root@vgpu ~]#

6.2.4. Getting vGPU Details

To get detailed information about all the vGPUs on the platform, run nvidia-smi vgpu with the –q or --query option.

To limit the information retrieved to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -q -i 1
GPU 00000000:C1:00.0
    Active vGPUs                          : 1
    vGPU ID                               : 3251634327
        VM ID                             : 2103066
        VM Name                           : Win11SV2_View87
        vGPU Name                         : GRID T4-4Q
        vGPU Type                         : 232
        vGPU UUID                         : afdcf724-1dd2-11b2-8534-624f22674b66
        Guest Driver Version              : 527.15
        License Status                    : Licensed (Expiry: 2022-11-23 5:2:12 GMT)
        GPU Instance ID                   : N/A
        Accounting Mode                   : Disabled
        ECC Mode                          : Enabled
        Accounting Buffer Size            : 4000
        Frame Rate Limit                  : 60 FPS
        PCI
            Bus Id                        : 00000000:02:04.0
        FB Memory Usage
            Total                         : 4096 MiB
            Used                          : 641 MiB
            Free                          : 3455 MiB
        Utilization
            Gpu                           : 0 %
            Memory                        : 0 %
            Encoder                       : 0 %
            Decoder                       : 0 %
        Encoder Stats
            Active Sessions               : 0
            Average FPS                   : 0
            Average Latency               : 0
        FBC Stats
            Active Sessions               : 0
            Average FPS                   : 0
            Average Latency               : 0
[root@vgpu ~]#

6.2.5. Monitoring vGPU engine usage

To monitor vGPU engine usage across multiple vGPUs, run nvidia-smi vgpu with the –u or --utilization option.

For each vGPU, the usage statistics in the following table are reported once every second. The table also shows the name of the column in the command output under which each statistic is reported.

Statistic Column
3D/Compute sm
Memory controller bandwidth mem
Video encoder enc
Video decoder dec

Each reported percentage is the percentage of the physical GPU’s capacity that a vGPU is using. For example, a vGPU that uses 20% of the GPU’s graphics engine’s capacity will report 20%.

To modify the reporting frequency, use the –l or --loop option.

To limit monitoring to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -u
# gpu    vgpu    sm   mem   enc   dec
# Idx      Id     %     %     %     %
    0   11924     6     3     0     0
    1   11903     8     3     0     0
    2   11908    10     4     0     0
    3       -     -     -     -     -
    4       -     -     -     -     -
    5       -     -     -     -     -
    0   11924     6     3     0     0
    1   11903     9     3     0     0
    2   11908    10     4     0     0
    3       -     -     -     -     -
    4       -     -     -     -     -
    5       -     -     -     -     -
    0   11924     6     3     0     0
    1   11903     8     3     0     0
    2   11908    10     4     0     0
    3       -     -     -     -     -
    4       -     -     -     -     -
    5       -     -     -     -     -
^C[root@vgpu ~]#

6.2.6. Monitoring vGPU engine usage by applications

To monitor vGPU engine usage by applications across multiple vGPUs, run nvidia-smi vgpu with the –p option.

For each application on each vGPU, the usage statistics in the following table are reported once every second. Each application is identified by its process ID and process name. The table also shows the name of the column in the command output under which each statistic is reported.

Statistic Column
3D/Compute sm
Memory controller bandwidth mem
Video encoder enc
Video decoder dec

Each reported percentage is the percentage of the physical GPU’s capacity used by an application running on a vGPU that resides on the physical GPU. For example, an application that uses 20% of the GPU’s graphics engine’s capacity will report 20%.

To modify the reporting frequency, use the –l or --loop option.

To limit monitoring to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -p
# GPU    vGPU process        process   sm  mem  enc  dec
# Idx      Id      Id           name    %    %    %    %
    0   38127    1528        dwm.exe    0    0    0    0
    1   37408    4232  DolphinVS.exe   32   25    0    0
    1  257869    4432    FurMark.exe   16   12    0    0
    1  257969    4552    FurMark.exe   48   37    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1   37408    4232  DolphinVS.exe   16   12    0    0
    1  257911     656  DolphinVS.exe   32   24    0    0
    1  257969    4552    FurMark.exe   48   37    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1  257869    4432    FurMark.exe   38   30    0    0
    1  257911     656  DolphinVS.exe   19   14    0    0
    1  257969    4552    FurMark.exe   38   30    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1  257848    3220    Balls64.exe   16   12    0    0
    1  257869    4432    FurMark.exe   16   12    0    0
    1  257911     656  DolphinVS.exe   16   12    0    0
    1  257969    4552    FurMark.exe   48   37    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1  257911     656  DolphinVS.exe   32   25    0    0
    1  257969    4552    FurMark.exe   64   50    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1   37408    4232  DolphinVS.exe   16   12    0    0
    1  257911     656  DolphinVS.exe   16   12    0    0
    1  257969    4552    FurMark.exe   64   49    0    0
    0   38127    1528        dwm.exe    0    0    0    0
    1   37408    4232  DolphinVS.exe   16   12    0    0
    1  257869    4432    FurMark.exe   16   12    0    0
    1  257969    4552    FurMark.exe   64   49    0    0
[root@vgpu ~]#

6.2.7. Monitoring Encoder Sessions

Note: Encoder sessions can be monitored only for vGPUs assigned to Windows VMs. No encoder session statistics are reported for vGPUs assigned to Linux VMs.

To monitor the encoder sessions for processes running on multiple vGPUs, run nvidia-smi vgpu with the –es or --encodersessions option.

For each encoder session, the following statistics are reported once every second:

  • GPU ID
  • vGPU ID
  • Encoder session ID
  • PID of the process in the VM that created the encoder session
  • Codec type, for example, H.264 or H.265
  • Encode horizontal resolution
  • Encode vertical resolution
  • One-second trailing average encoded FPS
  • One-second trailing average encode latency in microseconds

To modify the reporting frequency, use the –l or --loop option.

To limit monitoring to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -es
# GPU    vGPU Session Process   Codec       H       V Average     Average
# Idx      Id      Id      Id    Type     Res     Res     FPS Latency(us)
    1   21211       2    2308   H.264    1920    1080     424        1977
    1   21206       3    2424   H.264    1920    1080       0           0
    1   22011       1    3676   H.264    1920    1080     374        1589
    1   21211       2    2308   H.264    1920    1080     360         807
    1   21206       3    2424   H.264    1920    1080     325        1474
    1   22011       1    3676   H.264    1920    1080     313        1005
    1   21211       2    2308   H.264    1920    1080     329        1732
    1   21206       3    2424   H.264    1920    1080     352        1415
    1   22011       1    3676   H.264    1920    1080     434        1894
    1   21211       2    2308   H.264    1920    1080     362        1818
    1   21206       3    2424   H.264    1920    1080     296        1072
    1   22011       1    3676   H.264    1920    1080     416        1994
    1   21211       2    2308   H.264    1920    1080     444        1912
    1   21206       3    2424   H.264    1920    1080     330        1261
    1   22011       1    3676   H.264    1920    1080     436        1644
    1   21211       2    2308   H.264    1920    1080     344        1500
    1   21206       3    2424   H.264    1920    1080     393        1727
    1   22011       1    3676   H.264    1920    1080     364        1945
    1   21211       2    2308   H.264    1920    1080     555        1653
    1   21206       3    2424   H.264    1920    1080     295         925
    1   22011       1    3676   H.264    1920    1080     372        1869
    1   21211       2    2308   H.264    1920    1080     326        2206
    1   21206       3    2424   H.264    1920    1080     318        1366
    1   22011       1    3676   H.264    1920    1080     464        2015
    1   21211       2    2308   H.264    1920    1080     305        1167
    1   21206       3    2424   H.264    1920    1080     445        1892
    1   22011       1    3676   H.264    1920    1080     361         906
    1   21211       2    2308   H.264    1920    1080     353        1436
    1   21206       3    2424   H.264    1920    1080     354        1798
    1   22011       1    3676   H.264    1920    1080     373        1310 
^C[root@vgpu ~]#

6.2.8. Listing Supported vGPU Types

To list the virtual GPU types that the GPUs in the system support, run nvidia-smi vgpu with the –s or --supported option.

To limit the retrieved information to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -s -i 0
GPU 0000:83:00.0
    GRID M60-0B
    GRID M60-0Q
    GRID M60-1A
    GRID M60-1B
    GRID M60-1Q
    GRID M60-2A
    GRID M60-2Q
    GRID M60-4A
    GRID M60-4Q
    GRID M60-8A
    GRID M60-8Q
[root@vgpu ~]#

To view detailed information about the supported vGPU types, add the –v or --verbose option:

[root@vgpu ~]# nvidia-smi vgpu -s -i 0 -v   | less
GPU 00000000:40:00.0
    vGPU Type ID                          : 0xc
        Name                              : GRID M60-0Q
        Class                             : Quadro
        GPU Instance Profile ID           : N/A
        Max Instances                     : 16
        Max Instances Per VM              : 1
        Multi vGPU Exclusive              : False
        vGPU Exclusive Type               : False
        vGPU Exclusive Size               : False
        Device ID                         : 0x13f210de
        Sub System ID                     : 0x13f2114c
        FB Memory                         : 512 MiB
        Display Heads                     : 2
        Maximum X Resolution              : 2560
        Maximum Y Resolution              : 1600
        Frame Rate Limit                  : 60 FPS
        GRID License                      : Quadro-Virtual-DWS,5.0;GRID-Virtual-WS,2.0;GRID-Virtual-WS-Ext,2.0
    vGPU Type ID                          : 0xf
        Name                              : GRID M60-1Q
        Class                             : Quadro
        GPU Instance Profile ID           : N/A
        Max Instances                     : 8
        Max Instances Per VM              : 1
        Multi vGPU Exclusive              : False
        vGPU Exclusive Type               : False
        vGPU Exclusive Size               : False
        Device ID                         : 0x13f210de
        Sub System ID                     : 0x13f2114d
        FB Memory                         : 1024 MiB
        Display Heads                     : 4
        Maximum X Resolution              : 5120
        Maximum Y Resolution              : 2880
        Frame Rate Limit                  : 60 FPS
        GRID License                      : Quadro-Virtual-DWS,5.0;GRID-Virtual-WS,2.0;GRID-Virtual-WS-Ext,2.0
    vGPU Type ID                          : 0x12
        Name                              : GRID M60-2Q
        Class                             : Quadro
        GPU Instance Profile ID           : N/A
        Max Instances                     : 4
        Max Instances Per VM              : 1
        Multi vGPU Exclusive              : False
        vGPU Exclusive Type               : False
        vGPU Exclusive Size               : False
…
[root@vgpu ~]#

6.2.9. Listing the vGPU Types that Can Currently Be Created

To list the virtual GPU types that can currently be created on GPUs in the system, run nvidia-smi vgpu with the –c or --creatable option.

This property is a dynamic property that varies for each GPU depending on whether MIG mode is enabled for the GPU.

  • If MIG mode is not enabled for the GPU, or if the GPU does not support MIG, this property reflects the number and type of vGPUs that are already running on the GPU.
    • If no vGPUs are running on the GPU, all vGPU types that the GPU supports are listed.
    • If one or more vGPUs are running on the GPU, but the GPU is not fully loaded, only the type of the vGPUs that are already running is listed.
    • If the GPU is fully loaded, no vGPU types are listed.
  • If MIG mode is enabled for the GPU, the result reflects the number and type of GPU instances on which no vGPUs are already running.
    • If no GPU instances have been created, no vGPU types are listed.
    • If GPU instances have been created, only the vGPU types that correspond to GPU instances on which no vGPU is running are listed.
    • If a vGPU is running on every GPU instance, no vGPU types are listed.

To limit the retrieved information to a subset of the GPUs on the platform, use the –i or --id option to select one or more vGPUs.

[root@vgpu ~]# nvidia-smi vgpu -c -i 0
GPU 0000:83:00.0
    GRID M60-2Q
[root@vgpu ~]#

To view detailed information about the vGPU types that can currently be created, add the –v or --verbose option.

6.3. Monitoring GPU Performance from a Guest VM

You can use monitoring tools within an individual guest VM to monitor the performance of vGPUs or pass-through GPUs that are assigned to the VM. The scope of these tools is limited to the guest VM within which you use them. You cannot use monitoring tools within an individual guest VM to monitor any other GPUs in the platform.

For a vGPU, only these metrics are reported in a guest VM:

  • 3D/Compute
  • Memory controller
  • Video encoder
  • Video decoder
  • Frame buffer usage

Other metrics normally present in a GPU are not applicable to a vGPU and are reported as zero or N/A, depending on the tool that you are using.

6.3.1. Using nvidia-smi to Monitor GPU Performance from a Guest VM

In guest VMs, you can use the nvidia-smi command to retrieve statistics for the total usage by all applications running in the VM and usage by individual applications of the following resources:

  • GPU
  • Video encoder
  • Video decoder
  • Frame buffer

To use nvidia-smi to retrieve statistics for the total resource usage by all applications running in the VM, run the following command:

nvidia-smi dmon

To use nvidia-smi to retrieve statistics for resource usage by individual applications running in the VM, run the following command:

nvidia-smi pmon

7. Changing Scheduling Behavior for Time-Sliced vGPUs

NVIDIA GPUs implement a best effort vGPU scheduler that aims to balance performance across vGPUs. The best effort scheduler allows a vGPU to use GPU processing cycles that are not being used by other vGPUs. Under some circumstances, a VM running a graphics-intensive application may adversely affect the performance of graphics-light applications running in other VMs.

To address this issue with the best effort vGPU scheduler, NVIDIA GPUs additionally support equal share and fixed share vGPU schedulers. These schedulers impose a limit on GPU processing cycles used by a vGPU, which prevents graphics-intensive applications running in one VM from affecting the performance of graphics-light applications running in other VMs. On GPUs that support multiple vGPU schedulers, you can select the vGPU scheduler to use. You can also set the length of the time slice for the equal share and fixed share vGPU schedulers.

Note: If you use the equal share or fixed share vGPU scheduler, the frame-rate limiter (FRL) is disabled.

The best effort scheduler is the default scheduler for all supported GPU architectures.

7.1. Scheduling Policies for Time-Sliced vGPUs

In addition to the default best effort scheduler, GPUs based on NVIDIA GPU architectures after the Maxwell architecture support equal share and fixed share vGPU schedulers.

Equal share scheduler
The physical GPU is shared equally amongst the running vGPUs that reside on it. As vGPUs are added to or removed from a GPU, the share of the GPU's processing cycles allocated to each vGPU changes accordingly. As a result, the performance of a vGPU may increase as other vGPUs on the same GPU are stopped, or decrease as other vGPUs are started on the same GPU.
Fixed share scheduler
Each vGPU is given a fixed share of the physical GPU's processing cycles, the amount of which depends on the vGPU type, which in turn determines the maximum number of vGPUs per physical GPU. For example, the maximum number of T4-4C vGPUs per physical GPU is 4. When the scheduling policy is fixed share, each T4-4C vGPU is given one quarter, or 25%, the physical GPU's processing cycles. As vGPUs are added to or removed from a GPU, the share of the GPU's processing cycles allocated to each vGPU remains constant. As a result, the performance of a vGPU remains unchanged as other vGPUs are stopped or started on the same GPU.

By default, these schedulers impose a strict round-robin scheduling policy. When this policy is enforced, the schedulers maintain scheduling fairness by adjusting the time slice for each VM that is configured with NVIDIA vGPU. The strict round-robin scheduling policy ensures more consistent scheduling of the work for VMs that are configured with NVIDIA vGPU and restricts the impact of GPU-intensive applications running in one VM on applications running in other VMs.

Instead of a strict round-robin scheduling policy, you can ensure scheduling fairness by scheduling the work for the vGPU that has spent the least amount of time in the scheduled state. This behavior was the default scheduling behavior in NVIDIA AI Enterprise releases before 15.0. To control whether the schedulers impose a strict round-robin scheduling policy, use the RmPVMRL registry key.

When a strict round-robin scheduling policy is enforced, the adjustment to the time slice is based on the scheduling frequency and an averaging factor.

Scheduling frequency
The number of times per second that work for a specific vGPU is scheduled. The default scheduling frequency depends on the number of vGPUs that reside on the physical GPU:
  • If fewer than eight vGPUs reside on the physical GPU, the default is 480 Hz.
  • If eight or more vGPUs reside on the physical GPU, the default is 960 Hz.
Averaging factor
A number that determines the moving average of time-slice overshoots accrued for each vGPU. This average controls the strictness with which the scheduling frequency is enforced. A high value for the averaging factor enforces the scheduling frequency less strictly than a low value.

Deviations from the specified scheduling frequency occur because the actual amount of time that a scheduler allocates to a VM might exceed, or overshoot, the time slice specified for the VM. A scheduler enforces the scheduling frequency by shortening the next time slice for each vGPU VM to compensate for the accrued overshoot time of the VM.

To calculate the amount by which to shorten the next time slice for a vGPU VM, the scheduler maintains a running total of the accrued overshoot time for each vGPU VM. This amount is equal to the running total divided by the averaging factor that you specify. The calculated amount is also subtracted from the accrued overshoot time. A high value for the averaging factor enforces the scheduling frequency less strictly by spreading the compensation for the accrued overshoot time over a longer period.

To set the scheduling frequency and averaging factor, use the RmPVMRL registry key.

7.2. Scheduler Time Slice for Time-Sliced vGPUs

When multiple VMs access the vGPUs on a single GPU, the GPU performs the work for each VM serially. The vGPU scheduler time slice represents the amount of time that the work of a VM is allowed to run on the GPU before it is preempted and the work of the next VM is performed.

For the equal share and fixed share vGPU schedulers, you can set the length of the time slice. The length of the time slice affects latency and throughput. The optimal length of the time slice depends the workload that the GPU is handling.

  • For workloads that require low latency, a shorter time slice is optimal. Typically, these workloads are applications that must generate output at a fixed interval, such as graphics applications that generate output at a frame rate of 60 FPS. These workloads are sensitive to latency and should be allowed to run at least once per interval. A shorter time slice reduces latency and improves responsiveness by causing the scheduler to switch more frequently between VMs.

  • For workloads that require maximum throughput, a longer time slice is optimal. Typically, these workloads are applications that must complete their work as quickly as possible and do not require responsiveness, such as CUDA applications. A longer time slice increases throughput by preventing frequent switching between VMs.

7.3. RmPVMRL Registry Key

The RmPVMRL registry key controls the scheduling behavior for NVIDIA vGPUs by setting the scheduling policy, the averaging factor and scheduling frequency for schedulers with a strict round-robin scheduling policy, and the length of the time slice for schedulers without a strict round-robin scheduling policy.

Note: You can change the vGPU scheduling behavior only on GPUs that support multiple vGPU schedulers, that is, GPUs based on NVIDIA GPU architectures after the Maxwell architecture.

Type

Dword

Contents

Value Meaning
0x00 (default) Best effort scheduler
0x01

Equal share scheduler with a strict round-robin scheduling policy and the default time slice length, scheduling frequency, and averaging factor

0x03

Equal share scheduler without a strict round-robin scheduling policy and the default time slice length

0xAAFFF001 Equal share scheduler with a strict round-robin scheduling policy and a user-defined averaging factor AA and a user-defined scheduling frequency FFF
0x00TT0003 Equal share scheduler without a strict round-robin scheduling policy and with a user-defined time slice length TT
0x11

Fixed share scheduler with a strict round-robin scheduling policy and the default time slice length, scheduling frequency, and averaging factor

0x13

Fixed share scheduler without a strict round-robin scheduling policy and with the default time slice length

0xAAFFF011 Fixed share scheduler with a strict round-robin scheduling policy and a user-defined averaging factor AA and a user-defined scheduling frequency FFF
0x00TT0013 Fixed share scheduler without a strict round-robin scheduling policy and with a user-defined time slice length TT

The default time slice length depends on the maximum number of vGPUs per physical GPU allowed for the vGPU type.

Maximum Number of vGPUs Default Time Slice Length
Less than or equal to 8 2 ms
Greater than 8 1 ms
AA
Two hexadecimal digits in the range 0x01 to 0x3C that set the averaging factor for the equal share and fixed share schedulers with a strict round-robin scheduling policy.

If AA is 0x01, the compensation for the accrued overshoot time is applied in a single time slice.

If AA is 0x3C, the compensation for the accrued overshoot time is spread over 60 (0x3C) time slices.

If AA is outside the range 0x01 to 0x3C, the default value of 33 is used.

FFF
Three hexadecimal digits in the range 0x3F to 0x3C0 that set the scheduling frequency for the equal share and fixed share schedulers with a strict round-robin scheduling policy. The time slice is inverse of scheduling frequency.

If FFF is 000 or outside the range 0x3F to 0x3C0, the scheduling frequency is set to the default scheduling frequency for the vGPU type.

TT
Two hexadecimal digits in the range 01 to 1E that set the length of the time slice in milliseconds (ms) for the equal share and fixed share schedulers. The minimum length is 1 ms and the maximum length is 30 ms.

If TT is 00, the length is set to the default length for the vGPU type.

If TT is greater than 1E, the length is set to 30 ms.

Examples

This example sets the vGPU scheduler to equal share scheduler with a strict round-robin scheduling policy and the default time slice length, scheduling frequency, and averaging factor.

RmPVMRL=0x01

This example sets the vGPU scheduler to equal share scheduler without a strict round-robin scheduling policy and with a time slice that is 3 ms long.

RmPVMRL=0x00030003

This example sets the vGPU scheduler to fixed share scheduler with a strict round-robin scheduling policy and the default time slice length, scheduling frequency, and averaging factor.

RmPVMRL=0x11

This example sets the vGPU scheduler to fixed share scheduler without a strict round-robin scheduling policy and with a time slice that is 24 (0x18) ms long.

RmPVMRL=0x00180011

This example sets the vGPU scheduler to equal share scheduler with a strict round-robin scheduling policy, an averaging factor of 60 (0x3C), and a scheduling frequency of 960 (0x3C0) Hz.

RmPVMRL=0x3c3c0001

This example sets the vGPU scheduler to fixed share scheduler with a strict round-robin scheduling policy, an averaging factor of 60 (0x3C), and a scheduling frequency of 960 (0x3C0) Hz.

RmPVMRL=0x3c3c0011

7.4. Getting the Current Time-Sliced vGPU Scheduling Policy for All GPUs

You can use the hypervisor's dmesg command to get the current time-sliced vGPU scheduling policy for all GPUs. Get this information before changing the scheduling behavior of one or more GPUs to determine if you need to change it or after changing it to confirm the change.

Perform this task in your hypervisor command shell.

  1. Open a command shell on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Use the dmesg command to display messages from the kernel that contain the strings NVRM and scheduler.
    $ dmesg | grep NVRM | grep scheduler

    The scheduling policy is indicated in these messages by the following strings:

    • BEST_EFFORT
    • EQUAL_SHARE
    • FIXED_SHARE

    If the scheduling policy is equal share or fixed share, the scheduler time slice in ms is also displayed.

    This example gets the scheduling policy of the GPUs in a system in which the policy of one GPU is set to best effort, one GPU is set to equal share, and one GPU is set to fixed share.

    $ dmesg | grep NVRM | grep scheduler
    2020-10-05T02:58:08.928Z cpu79:2100753)NVRM: GPU at 0000:3d:00.0 has software scheduler DISABLED with policy BEST_EFFORT.
    2020-10-05T02:58:09.818Z cpu79:2100753)NVRM: GPU at 0000:5e:00.0 has software scheduler ENABLED with policy EQUAL_SHARE.
    NVRM: Software scheduler timeslice set to 1 ms.
    2020-10-05T02:58:12.115Z cpu79:2100753)NVRM: GPU at 0000:88:00.0 has software scheduler ENABLED with policy FIXED_SHARE.
    NVRM: Software scheduler timeslice set to 1 ms.

7.5. Changing the Time-Sliced vGPU Scheduling Behavior for All GPUs

Perform this task in your hypervisor command shell.

  1. Open a command shell on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Set the RmPVMRL registry key to the value that sets the GPU scheduling policy and the length of the time slice that you want.
    • On Red Hat Enterprise Linux KVM, add the following entry to the /etc/modprobe.d/nvidia.conf file.

      options nvidia NVreg_RegistryDwords="RmPVMRL=value"

      If the /etc/modprobe.d/nvidia.conf file does not already exist, create it.

    • On VMware vSphere, use the esxcli set command.

      # esxcli system module parameters set -m nvidia -p "NVreg_RegistryDwords=RmPVMRL=value"
    value

    The value that sets the GPU scheduling policy and the length of the time slice that you want, for example:

    0x01
    Sets the vGPU scheduling policy to equal share scheduler with the default time slice length.
    0x00030001
    Sets the GPU scheduling policy to equal share scheduler with a time slice that is 3 ms long.
    0x11
    Sets the vGPU scheduling policy to fixed share scheduler with the default time slice length.
    0x00180011
    Sets the GPU scheduling policy to fixed share scheduler with a time slice that is 24 (0x18) ms long.

    For all supported values, see RmPVMRL Registry Key.

  3. Reboot your hypervisor host machine.
Confirm that the scheduling behavior was changed as required as explained in Getting the Current Time-Sliced vGPU Scheduling Policy for All GPUs.

7.6. Changing the Time-Sliced vGPU Scheduling Behavior for Select GPUs

Perform this task in your hypervisor command shell.

  1. Open a command shell on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Use the lspci command to obtain the PCI domain and bus/device/function (BDF) of each GPU for which you want to change the scheduling behavior.
    • On Red Hat Enterprise Linux KVM, add the -D option to display the PCI domain and the -d 10de: option to display information only for NVIDIA GPUs.
      # lspci -D -d 10de:
    • On VMware vSphere, pipe the output of lspci to the grep command to display information only for NVIDIA GPUs.
      # lspci | grep NVIDIA

    The NVIDIA GPU listed in this example has the PCI domain 0000 and BDF 86:00.0.

    0000:86:00.0 3D controller: NVIDIA Corporation GP104GL [Tesla P4] (rev a1)
  3. Use the module parameter NVreg_RegistryDwordsPerDevice to set the pci and RmPVMRL registry keys for each GPU.
    • On Red Hat Enterprise Linux KVM, add the following entry to the /etc/modprobe.d/nvidia.conf file.

      options nvidia NVreg_RegistryDwordsPerDevice="pci=pci-domain:pci-bdf;RmPVMRL=value
      [;pci=pci-domain:pci-bdf;RmPVMRL=value...]"

      If the /etc/modprobe.d/nvidia.conf file does not already exist, create it.

    • On VMware vSphere, use the esxcli set command.

      # esxcli system module parameters set -m nvidia \
      -p "NVreg_RegistryDwordsPerDevice=pci=pci-domain:pci-bdf;RmPVMRL=value\
      [;pci=pci-domain:pci-bdf;RmPVMRL=value...]"

    For each GPU, provide the following information:

    pci-domain
    The PCI domain of the GPU.
    pci-bdf
    The PCI device BDF of the GPU.
    value

    The value that sets the GPU scheduling policy and the length of the time slice that you want, for example:

    0x01
    Sets the GPU scheduling policy to equal share scheduler with the default time slice length.
    0x00030001
    Sets the GPU scheduling policy to equal share scheduler with a time slice that is 3 ms long.
    0x11
    Sets the GPU scheduling policy to fixed share scheduler with the default time slice length.
    0x00180011
    Sets the GPU scheduling policy to fixed share scheduler with a time slice that is 24 (0x18) ms long.

    For all supported values, see RmPVMRL Registry Key.

    This example adds an entry to the /etc/modprobe.d/nvidia.conf file to change the scheduling behavior of a single GPU. The entry sets the GPU scheduling policy of the GPU at PCI domain 0000 and BDF 86:00.0 to fixed share scheduler with the default time slice length.

    options nvidia NVreg_RegistryDwordsPerDevice=
    "pci=0000:86:00.0;RmPVMRL=0x11"

    This example adds an entry to the /etc/modprobe.d/nvidia.conf file to change the scheduling behavior of a single GPU. The entry sets the scheduling policy of the GPU at PCI domain 0000 and BDF 86:00.0 to fixed share scheduler with a time slice that is 24 (0x18) ms long.

    options nvidia NVreg_RegistryDwordsPerDevice=
    "pci=0000:86:00.0;RmPVMRL=0x00180011"

    This example changes the scheduling behavior of a single GPU on a hypervisor host that is running VMware vSphere. The command sets the scheduling policy of the GPU at PCI domain 0000 and BDF 15:00.0 to fixed share scheduler with the default time slice length.

    # esxcli system module parameters set -m nvidia -p \
    "NVreg_RegistryDwordsPerDevice=pci=0000:15:00.0;RmPVMRL=0x11[;pci=0000:15:00.0;RmPVMRL=0x11]"

    This example changes the scheduling behavior of a single GPU on a hypervisor host that is running VMware vSphere. The command sets the scheduling policy of the GPU at PCI domain 0000 and BDF 15:00.0 to fixed share scheduler with a time slice that is 24 (0x18) ms long.

    # esxcli system module parameters set -m nvidia -p \
    "NVreg_RegistryDwordsPerDevice=pci=0000:15:00.0;RmPVMRL=0x11[;pci=0000:15:00.0;RmPVMRL=0x00180011]"
  4. Reboot your hypervisor host machine.
Confirm that the scheduling behavior was changed as required as explained in Getting the Current Time-Sliced vGPU Scheduling Policy for All GPUs.

7.7. Restoring Default Time-Sliced vGPU Scheduler Settings

Perform this task in your hypervisor command shell.

  1. Open a command shell on your hypervisor host machine. On all supported hypervisors, you can use secure shell (SSH) for this purpose. Individual hypervisors may provide additional means for logging in. For details, refer to the documentation for your hypervisor.
  2. Unset the RmPVMRL registry key.
    • On Red Hat Enterprise Linux KVM, comment out the entries in the /etc/modprobe.d/nvidia.conf file that set RmPVMRL by prefixing each entry with the # character.

    • On VMware vSphere, set the module parameter to an empty string.

      # esxcli system module parameters set -m nvidia -p "module-parameter="
      module-parameter
      The module parameter to set, which depends on whether the scheduling behavior was changed for all GPUs or select GPUs:
      • For all GPUs, set the NVreg_RegistryDwords module parameter.
      • For select GPUs, set the NVreg_RegistryDwordsPerDevice module parameter.

      For example, to restore default vGPU scheduler settings after they were changed for all GPUs, enter this command:

      # esxcli system module parameters set -m nvidia -p "NVreg_RegistryDwords="
  3. Reboot your hypervisor host machine.

8. Troubleshooting

This chapter describes basic troubleshooting steps for NVIDIA vGPU and how to collect debug information when filing a bug report.

8.1. Known issues

Before troubleshooting or filing a bug report, review the release notes that accompany each driver release, for information about known issues with the current release, and potential workarounds.

8.2. Troubleshooting steps

If a vGPU-enabled VM fails to start, or doesn’t display any output when it does start, follow these steps to narrow down the probable cause.

8.2.1. Verifying the NVIDIA Kernel Driver Is Loaded

  1. Run the vmkload_mod command.
    [root@esxi:~] vmkload_mod -l | grep nvidia
    nvidia                   5    8420
  2. If the nvidia driver is not listed in the output, check dmesg for any load-time errors reported by the driver (see Examining NVIDIA kernel driver output).

8.2.2. Verifying that nvidia-smi works

If the NVIDIA kernel driver is correctly loaded on the physical GPU, run nvidia-smi and verify that all physical GPUs not currently being used for GPU pass-through are listed in the output. For details on expected output, see NVIDIA System Management Interface nvidia-smi.

If nvidia-smi fails to report the expected output, check dmesg for NVIDIA kernel driver messages.

8.2.3. Examining NVIDIA kernel driver output

Information and debug messages from the NVIDIA kernel driver are logged in kernel logs, prefixed with NVRM or nvidia.

Run dmesg and check for the NVRM and nvidia prefixes:

[root@xenserver ~]# dmesg | grep -E "NVRM|nvidia"
[   22.054928] nvidia: module license 'NVIDIA' taints kernel.
[   22.390414] NVRM: loading
[   22.829226] nvidia 0000:04:00.0: enabling device (0000 -> 0003)
[   22.829236] nvidia 0000:04:00.0: PCI INT A -> GSI 32 (level, low) -> IRQ 32
[   22.829240] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
[   22.829241] NVRM: BAR0 is 0M @ 0x0 (PCI:0000:00:04.0)
[   22.829243] NVRM: The system BIOS may have misconfigured your GPU.

8.2.4. Examining NVIDIA Virtual GPU Manager Messages

Information and debug messages from the NVIDIA Virtual GPU Manager are logged to the hypervisor’s log files, prefixed with vmiop.

8.2.4.1. Examining VMware vSphere vGPU Manager Messages

For VMware vSphere, NVIDIA Virtual GPU Manager messages are written to the vmware.log file in the guest VM’s storage directory.

Look in the vmware.log file for the vmiop prefix:

[root@esxi:~] grep vmiop /vmfs/volumes/datastore1/win7-vgpu-test1/vmware.log
2023-08-17T14:02:21.275Z| vmx| I120: DICT   pciPassthru0.virtualDev = "vmiop"
2023-08-17T14:02:21.344Z| vmx| I120: GetPluginPath testing /usr/lib64/vmware/plugin/libvmx-vmiop.so
2023-08-17T14:02:21.344Z| vmx| I120: PluginLdr_LoadShared: Loaded shared plugin libvmx-vmiop.so from /usr/lib64/vmware/plugin/libvmx-vmiop.so
2023-08-17T14:02:21.344Z| vmx| I120: VMIOP: Loaded plugin libvmx-vmiop.so:VMIOP_InitModule
2023-08-17T14:02:21.359Z| vmx| I120: VMIOP: Initializing plugin vmiop-display
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: gpu-pci-id : 0000:04:00.0
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: vgpu_type : quadro
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: Framebuffer: 0x74000000
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: Virtual Device Id: 0x11B0:0x101B
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: ######## vGPU Manager Information: ########
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: Driver Version: 525.125.03
2023-08-17T14:02:21.365Z| vmx| I120: vmiop_log: VGX Version: 15.3
2023-08-17T14:02:21.445Z| vmx| I120: vmiop_log: Init frame copy engine: syncing...
2023-08-17T14:02:37.031Z| vthread-12| I120: vmiop_log: ######## Guest NVIDIA Driver Information: ########
2023-08-17T14:02:37.031Z| vthread-12| I120: vmiop_log: Driver Version: 529.11
2023-08-17T14:02:37.031Z| vthread-12| I120: vmiop_log: VGX Version: 15.3
2023-08-17T14:02:37.093Z| vthread-12| I120: vmiop_log: Clearing BAR1 mapping
2023-08-21T23:39:55.726Z| vmx| I120: VMIOP: Shutting down plugin vmiop-display
[root@esxi:~]

8.3. Capturing configuration data by running nvidia-bug-report.sh

The nvidia-bug-report.sh script captures debug information into a gzip-compressed log file on the server.

Run nvidia-bug-report.sh from the VMware ESXi host shell.

9. Additional Information

Additional information about containers and resources that are available from the NVIDIA AI Enterprise private registry on NGC is available in the documentation for these entities.

Container or Resource Additional Information
TensorFlow
PyTorch PyTorch Release Notes
NVIDIA Triton Inference Server Triton Inference Server Documentation on Github
NVIDIA TensorRT NVIDIA TensorRT Documentation
NVIDIA RAPIDS RAPIDS Docs on the RAPIDS project site
NVIDIA RAPIDS Accelerator for Apache Spark RAPIDS Accelerator for Apache Spark Deployment Guide
NVIDIA Clara Parabricks NVIDIA Clara Parabricks Documentation
NVIDIA DeepStream
MONAI (Medical Open Network for Artificial Intelligence) Enterprise NVIDIA MONAI Toolkit Documentation
TAO Toolkit TAO Toolkit Documentation
NVIDIA GPU Operator NVIDIA GPU Operator Documentation
NVIDIA Network Operator NVIDIA Network Operator Documentation

A. Virtual GPU Types for Supported GPUs

NVIDIA vGPU is available as a licensed product on supported NVIDIA GPUs. For a list of recommended server platforms and supported GPUs, consult the release notes for supported hypervisors at NVIDIA AI Enterprise Documentation.

A.1. NVIDIA A800 PCIe 80GB and NVIDIA A800 PCIe 80GB Liquid Cooled Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

The virtual GPU types for the NVIDIA A800 PCIe 80GB and NVIDIA A800 PCIe 80GB liquid cooled are identical.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A800 PCIe 80GB and NVIDIA A800 PCIe 80GB Liquid Cooled

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A800D-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
A800D-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
A800D-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
A800D-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
A800D-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
A800D-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A800 PCIe 80GB and NVIDIA A800 PCIe 80GB Liquid Cooled

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A800D-80C Training Workloads 81920 1 1 4096×21601 1
A800D-40C Training Workloads 40960 2 2 4096×21601 1
A800D-20C Training Workloads 20480 4 4 4096×21601 1
A800D-16C Inference Workloads 16384 5 5 4096×21601 1
A800D-10C Training Workloads 10240 8 8 4096×21601 1
A800D-8C Training Workloads 8192 10 10 4096×21601 1
A800D-4C Inference Workloads 4096 20 20 4096×21601 1

A.2. NVIDIA A800 HGX Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A800 HGX 80GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A800DX-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
A800DX-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
A800DX-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
A800DX-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
A800DX-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
A800DX-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A800 HGX 80GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A800DX-80C Training Workloads 81920 1 1 4096×21601 1
A800DX-40C Training Workloads 40960 2 2 4096×21601 1
A800DX-20C Training Workloads 20480 4 4 4096×21601 1
A800DX-16C Inference Workloads 16384 5 5 4096×21601 1
A800DX-10C Training Workloads 10240 8 8 4096×21601 1
A800DX-8C Training Workloads 8192 10 10 4096×21601 1
A800DX-4C Inference Workloads 4096 20 20 4096×21601 1

A.3. NVIDIA A100 PCIe 40GB Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A100 PCIe 40GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A100-7-40C Training Workloads 40960 1 7 7 MIG 7g.40gb
A100-4-20C Training Workloads 20480 1 4 4 MIG 4g.20gb
A100-3-20C Training Workloads 20480 2 3 3 MIG 3g.20gb
A100-2-10C Training Workloads 10240 3 2 2 MIG 2g.10gb
A100-1-5C Inference Workloads 5120 7 1 1 MIG 1g.5gb
A100-1-5CME Inference Workloads 5120 1 1 1 MIG 1g.5gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A100 PCIe 40GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A100-40C Training Workloads 40960 1 1 4096×21601 1
A100-20C Training Workloads 20480 2 2 4096×21601 1
A100-10C Training Workloads 10240 4 4 4096×21601 1
A100-8C Training Workloads 8192 5 5 4096×21601 1
A100-5C Inference Workloads 5120 8 8 4096×21601 1
A100-4C Inference Workloads 4096 10 10 4096×21601 1

A.4. NVIDIA A100 HGX 40GB Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A100 HGX 40GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A100X-7-40C Training Workloads 40960 1 7 7 MIG 7g.40gb
A100X-4-20C Training Workloads 20480 1 4 4 MIG 4g.20gb
A100X-3-20C Training Workloads 20480 2 3 3 MIG 3g.20gb
A100X-2-10C Training Workloads 10240 3 2 2 MIG 2g.10gb
A100X-1-5C Inference Workloads 5120 7 1 1 MIG 1g.5gb
A100X-1-5CME Inference Workloads 5120 1 1 1 MIG 1g.5gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A100 HGX 40GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A100X-40C Training Workloads 40960 1 1 4096×21601 1
A100X-20C Training Workloads 20480 2 2 4096×21601 1
A100X-10C Training Workloads 10240 4 4 4096×21601 1
A100X-8C Training Workloads 8192 5 5 4096×21601 1
A100X-5C Inference Workloads 5120 8 8 4096×21601 1
A100X-4C Inference Workloads 4096 10 10 4096×21601 1

A.5. NVIDIA A100 PCIe 80GB, NVIDIA A100 PCIe 80GB Liquid Cooled and NVIDIA A100X Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

The virtual GPU types for the NVIDIA A100 PCIe 80GB, NVIDIA A100 PCIe 80GB liquid cooled and NVIDIA A100X GPUs are identical.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A100 PCIe 80GB, NVIDIA A100 PCIe 80GB Liquid Cooled and NVIDIA A100X

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A100D-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
A100D-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
A100D-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
A100D-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
A100D-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
A100D-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A100 PCIe 80GB, NVIDIA A100 PCIe 80GB Liquid Cooled and NVIDIA A100X

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A100D-80C Training Workloads 81920 1 1 4096×21601 1
A100D-40C Training Workloads 40960 2 2 4096×21601 1
A100D-20C Training Workloads 20480 4 4 4096×21601 1
A100D-16C Inference Workloads 16384 5 5 4096×21601 1
A100D-10C Training Workloads 10240 8 8 4096×21601 1
A100D-8C Training Workloads 8192 10 10 4096×21601 1
A100D-4C Inference Workloads 4096 20 20 4096×21601 1

A.6. NVIDIA A100 HGX 80GB Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A100 HGX 80GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A100DX-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
A100DX-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
A100DX-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
A100DX-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
A100DX-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
A100DX-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A100 HGX 80GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A100DX-80C Training Workloads 81920 1 1 4096×21601 1
A100DX-40C Training Workloads 40960 2 2 4096×21601 1
A100DX-20C Training Workloads 20480 4 4 4096×21601 1
A100DX-16C Inference Workloads 16384 5 5 4096×21601 1
A100DX-10C Training Workloads 10240 8 8 4096×21601 1
A100DX-8C Training Workloads 8192 10 10 4096×21601 1
A100DX-4C Inference Workloads 4096 20 20 4096×21601 1

A.7. NVIDIA A40 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA A40

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A40-48C Training Workloads 49152 1 1 4096×21601 1
A40-24C Training Workloads 24576 2 2 4096×21601 1
A40-16C Training Workloads 16384 3 3 4096×21601 1
A40-12C Training Workloads 12288 4 4 4096×21601 1
A40-8C Training Workloads 8192 6 6 4096×21601 1
A40-6C Training Workloads 6144 8 8 4096×21601 1
A40-4C Inference Workloads 4096 122 12 4096×21601 1

A.8. NVIDIA A30 and NVIDIA A30X Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

The virtual GPU types for the NVIDIA A30 and NVIDIA A30X GPUs are identical.

MIG-Backed C-Series Virtual GPU Types for NVIDIA A30 and NVIDIA A30X

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
A30-4-24C Training Workloads 24576 1 4 4 MIG 4g.24gb
A30-2-12C Training Workloads 12288 2 2 2 MIG 2g.12gb
A30-1-6C Inference Workloads 6144 4 1 1 MIG 1g.6gb
A30-1-6CME Inference Workloads 6144 1 1 1 MIG 1g.6gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA A30 and NVIDIA A30X

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A30-24C Training Workloads 24576 1 1 4096×21601 1
A30-12C Training Workloads 12288 2 2 4096×21601 1
A30-8C Training Workloads 8192 3 3 4096×21601 1
A30-6C Inference Workloads 6144 4 4 4096×21601 1
A30-4C Inference Workloads 4096 6 6 4096×21601 1

A.9. NVIDIA A16 Virtual GPU Types

Physical GPUs per board: 4

C-Series Virtual GPU Types for NVIDIA A16

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A16-16C Training Workloads 16384 1 4 4096×21601 1
A16-8C Training Workloads 8192 2 8 4096×21601 1
A16-4C Inference Workloads 4096 4 16 4096×21601 1

A.10. NVIDIA A10 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA A10

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
A10-24C Training Workloads 24576 1 1 4096×21601 1
A10-12C Training Workloads 12288 2 2 4096×21601 1
A10-8C Training Workloads 8192 3 3 4096×21601 1
A10-6C Training Workloads 6144 4 4 4096×21601 1
A10-4C Inference Workloads 4096 6 6 4096×21601 1

A.11. NVIDIA H100 PCIe 80GB Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA H100 PCIe 80GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
H100-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
H100-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
H100-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
H100-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
H100-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
H100-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA H100 PCIe 80GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
H100-80C Training Workloads 81920 1 1 4096×21601 1
H100-40C Training Workloads 40960 2 2 4096×21601 1
H100-20C Training Workloads 20480 4 4 4096×21601 1
H100-16C Inference Workloads 16384 5 5 4096×21601 1
H100-10C Training Workloads 10240 8 8 4096×21601 1
H100-8C Training Workloads 8192 10 10 4096×21601 1
H100-5C Inference Workloads 5120 16 16 4096×21601 1
H100-4C Inference Workloads 4096 20 20 4096×21601 1

A.12. NVIDIA H800 PCIe 80GB Virtual GPU Types

Physical GPUs per board: 1

This GPU supports MIG-backed virtual GPUs and time-sliced virtual GPUs.

MIG-Backed C-Series Virtual GPU Types for NVIDIA H800 PCIe 80GB

Required license edition: vCS or vWS

For details of GPU instance profiles, see NVIDIA Multi-Instance GPU User Guide.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Slices per vGPU Compute Instances per vGPU Corresponding GPU Instance Profile
H800-7-80C Training Workloads 81920 1 7 7 MIG 7g.80gb
H800-4-40C Training Workloads 40960 1 4 4 MIG 4g.40gb
H800-3-40C Training Workloads 40960 2 3 3 MIG 3g.40gb
H800-2-20C Training Workloads 20480 3 2 2 MIG 2g.20gb
H800-1-10C Training Workloads 10240 7 1 1 MIG 1g.10gb
H800-1-10CME Training Workloads 10240 1 1 1 MIG 1g.10gb+me

Time-Sliced C-Series Virtual GPU Types for NVIDIA H800 PCIe 80GB

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
H800-80C Training Workloads 81920 1 1 4096×21601 1
H800-40C Training Workloads 40960 2 2 4096×21601 1
H800-20C Training Workloads 20480 4 4 4096×21601 1
H800-16C Inference Workloads 16384 5 5 4096×21601 1
H800-10C Training Workloads 10240 8 8 4096×21601 1
H800-8C Training Workloads 8192 10 10 4096×21601 1
H800-5C Inference Workloads 5120 16 16 4096×21601 1
H800-4C Inference Workloads 4096 20 20 4096×21601 1

A.13. NVIDIA L40 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA L40

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
L40-48C Training Workloads 49152 1 1 4096×21601 1
L40-24C Training Workloads 24576 2 2 4096×21601 1
L40-16C Training Workloads 16384 3 3 4096×21601 1
L40-12C Training Workloads 12288 4 4 4096×21601 1
L40-8C Training Workloads 8192 6 6 4096×21601 1
L40-6C Training Workloads 6144 8 8 4096×21601 1
L40-4C Inference Workloads 4096 122 12 4096×21601 1

A.14. NVIDIA L4 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA L4

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
L4-24C Training Workloads 24576 1 1 4096×21601 1
L4-12C Training Workloads 12288 2 2 4096×21601 1
L4-8C Training Workloads 8192 3 3 4096×21601 1
L4-6C Training Workloads 6144 4 4 4096×21601 1
L4-4C Inference Workloads 4096 6 6 4096×21601 1

A.15. NVIDIA RTX 6000 Ada Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA RTX 6000 Ada

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
RTX 6000 Ada-48C Training Workloads 49152 1 1 4096×21601 1
RTX 6000 Ada-24C Training Workloads 24576 2 2 4096×21601 1
RTX 6000 Ada-16C Training Workloads 16384 3 3 4096×21601 1
RTX 6000 Ada-12C Training Workloads 12288 4 4 4096×21601 1
RTX 6000 Ada-8C Training Workloads 8192 6 6 4096×21601 1
RTX 6000 Ada-6C Training Workloads 6144 8 8 4096×21601 1
RTX 6000 Ada-4C Inference Workloads 4096 122 12 4096×21601 1

A.16. NVIDIA RTX A6000 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA RTX A6000

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
RTXA6000-48C Training Workloads 49152 1 1 4096×21601 1
RTXA6000-24C Training Workloads 24576 2 2 4096×21601 1
RTXA6000-16C Training Workloads 16384 3 3 4096×21601 1
RTXA6000-12C Training Workloads 12288 4 4 4096×21601 1
RTXA6000-8C Training Workloads 8192 6 6 4096×21601 1
RTXA6000-6C Training Workloads 6144 8 8 4096×21601 1
RTXA6000-4C Inference Workloads 4096 122 12 4096×21601 1

A.17. NVIDIA RTX A5000 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for NVIDIA RTX A5000

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
RTXA5000-24C Training Workloads 24576 1 1 4096×21601 1
RTXA5000-12C Training Workloads 12288 2 2 4096×21601 1
RTXA5000-8C Training Workloads 8192 3 3 4096×21601 1
RTXA5000-6C Training Workloads 6144 4 4 4096×21601 1
RTXA5000-4C Inference Workloads 4096 6 6 4096×21601 1

A.18. Tesla T4 Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for Tesla T4

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
T4-16C Training Workloads 16384 1 1 4096×21601 1
T4-8C Training Workloads 8192 2 2 4096×21601 1
T4-4C Inference Workloads 4096 4 4 4096×21601 1

A.19. Quadro RTX 8000 Passive Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for Quadro RTX 8000 Passive

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
RTX8000P-48C Training Workloads 49152 1 1 4096×21601 1
RTX8000P-24C Training Workloads 24576 2 2 4096×21601 1
RTX8000P-16C Training Workloads 16384 3 3 4096×21601 1
RTX8000P-12C Training Workloads 12288 4 4 4096×21601 1
RTX8000P-8C Training Workloads 8192 6 6 4096×21601 1
RTX8000P-6C Training Workloads 6144 8 8 4096×21601 1
RTX8000P-4C Inference Workloads 4096 82 8 4096×21601 1

A.20. Quadro RTX 6000 Passive Virtual GPU Types

Physical GPUs per board: 1

C-Series Virtual GPU Types for Quadro RTX 6000 Passive

Required license edition: vCS or vWS

These vGPU types support a single display with a fixed maximum resolution.

Virtual GPU Type Intended Use Case Frame Buffer (MB) Maximum vGPUs per GPU Maximum vGPUs per Board Maximum Display Resolution Virtual Displays per vGPU
RTX6000P-24C Training Workloads 24576 1 1 4096×21601 1
RTX6000P-12C Training Workloads 12288 2 2 4096×21601 1
RTX6000P-8C Training Workloads 8192 3 3 4096×21601 1
RTX6000P-6C Training Workloads 6144 4 4 4096×21601 1
RTX6000P-4C Inference Workloads 4096 6 6 4096×21601 1

Notices

Notice

This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a product. NVIDIA Corporation (“NVIDIA”) makes no representations or warranties, expressed or implied, as to the accuracy or completeness of the information contained in this document and assumes no responsibility for any errors contained herein. NVIDIA shall have no liability for the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.

NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any time without notice.

Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.

NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No contractual obligations are formed either directly or indirectly by this document.

NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such inclusion and/or use is at customer’s own risk.

NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product designs.

No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.

Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.

THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the product.

VESA DisplayPort

DisplayPort and DisplayPort Compliance Logo, DisplayPort Compliance Logo for Dual-mode Sources, and DisplayPort Compliance Logo for Active Cables are trademarks owned by the Video Electronics Standards Association in the United States and other countries.

HDMI

HDMI, the HDMI logo, and High-Definition Multimedia Interface are trademarks or registered trademarks of HDMI Licensing LLC.

OpenCL

OpenCL is a trademark of Apple Inc. used under license to the Khronos Group Inc.

Trademarks

NVIDIA, the NVIDIA logo, NVIDIA Maxwell, NVIDIA Pascal, NVIDIA Turing, NVIDIA Volta, Quadro, and Tesla are trademarks or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated.

1 C-series vGPU types are NVIDIA Virtual Compute Server vGPU types, which are optimized for compute-intensive workloads. As a result, they support only a single display head and do not provide Quadro graphics acceleration.
2 The maximum number of NVIDIA Virtual Compute Server vGPUs is limited to 12 vGPUs per physical GPU, irrespective of the available hardware resources of the physical GPU.