NVIDIA vGPU 14

It’s time to plan updating your NVIDIA Enterprise GPUs. NVIDIA vGPU Software 14 is now GA since February 2022.

NVIDIA vGPU software includes vWS, vCS, vPC, and vApps.

If you got any of following NVIDIA GPU’s: A100, A40, A30, A16, A10, A2, RTX A6000, RTX A5000, RTX8000, RTX6000, V100, T4, P100, P40, P6, P4, M60, M10, M6  If you are interested in a quick overview of which NVIDIA enterprise GPU is supporting which hypervisor, Guest os and remoting technology, I highly recommend you check out this link from NVIDIA that provides the NVIDIA vGPU software product support matrix. NVIDIA vGPU software 14 is supported until February 2023. NVIDIA vGPU software 14 is a Product Branch Support. I this article, I am also covering which Public Cloud instance is available with NVIDIA GPUs and which license is BYO or provided by the public cloud provider such as Azure, AWS, GCP, Alibaba. For a list of validated server platforms, refer to NVIDIA vGPU Certified Servers. Important note for EUC (Citrix/VMware customers):
  • NVIDIA vGPU 14.0 is supported with VMware Horizon 2111 (8.4), 2106 (8.3), 2013 (8.2), 2012 (8.1), 2006 (8.0), 7.13, 7.12, 7.11, 7.10, 7.9, 7.8, 7.7, 7.6, 7.5, 7.4, 7.3, 7.2, 7.1, 7.0
  • NVIDIA vGPU 14.0 is only supported with Citrix Virtual Apps & Desktops (LTSR) 7 1912, 7.15 (CR)  7 2112, 7 2109, 7 2106, 7 2103, 7 2012, 7 2009, 7 2003, 7 1909, 7 1906, 7 1903
This release includes the following software:
  • NVIDIA vGPU Manager version 510.47.03 for following hypervisors. (Citrix Hypervisor, VMware vSphere, RHEL KVM)
  • NVIDIA Windows driver version 511.65
  • NVIDIA Linux driver version 510.47.03

New Features in Release 14.0

  • Support for GPUDirect® technology on all C-series vGPUs on GPUs that support SR-IOV
  • Support for a mixture of time-sliced vGPUs of the same frame buffer size on the same GPU
  • Support for Tesla Compute Cluster (TCC) mode for Q-series vGPUs on Windows guest VMs
  • Support for GPU System Processor (GSP) in GPU pass through and bare-metal configurations on Linux with vCS
    Note: If you are using a product other than vCS, you must disable GSP as explained in Virtual GPU Software User Guide.
  • Enhanced NVIDIA CUDA Toolkit support:
    • NVIDIA CUDA Toolkit profilers can be enabled when unified memory is enabled.
    • Nsight Systems GPU context switch trace is supported.
  • Enhancements to the NVIDIA Management Library (NVML) to determine whether a vGPU type supports GPUDirect technology and peer-to-peer CUDA transfers over NVLink
  • Addition of RPM and Debian packages for the NVIDIA vGPU software graphics drivers for Linux
  • Security updates – see Security Bulletin: NVIDIA GPU Display Driver – February 2022, which is posted shortly after the release date of this software and is listed on the NVIDIA Product Security pages
  • Miscellaneous bug fixes

Hardware and Software Support Introduced in Release 14.0

  • Support for the following GPUs:
    • NVIDIA A2
    • NVIDIA A30X
    • NVIDIA A100X
  • Support for Red Hat Enterprise Linux with KVM hypervisor 8.5
  • Support for Red Hat Enterprise Linux 8.5 as a guest OS
  • Support for Debian 10 as a guest OS
  • Support for Citrix Virtual Apps and Desktops version 7 2112
  • Support for VMware Horizon 2111 (8.4)

Feature Support Withdrawn in Release 14.0

  • Red Hat Enterprise Linux with KVM hypervisor 8.1, 7.8, and 7.7 are no longer supported.
  • Red Hat Enterprise Linux 8.1 is no longer supported as a guest OS.
  • Red Hat Enterprise Linux 7.8 and 7.7 are no longer supported as a guest OS.
  • Windows Server 2012 R2 is no longer supported as a guest OS.

Features Deprecated in Release 14.0

The following table lists features that are deprecated in this release of NVIDIA vGPU software. Although the features remain available in this release, they might be withdrawn in a future release. In preparation for the possible removal of these features, use the preferred alternative listed in the table.

Deprecated Feature Preferred Alternative Additional Information
Legacy NVIDIA vGPU software license server NVIDIA License System NVIDIA Virtual GPU Software License Server End of Life Notice

Supported NVIDIA GPUs with vGPU 14.0

  • A100 HGX 80GB
  • A100 PCIe 80GB
  • A100X
  • A100 HGX 40GB
  • A100 PCIe 40GB
  • A40
  • A30
  • A30X
  • A16
  • A10
  • A2
  • RTX A6000
  • RTX A5000
  • RTX 8000
  • RTX 8000 passive
  • RTX 6000
  • RTX 6000 passive
  • V100
  • T4
  • P100
  • P40
  • P6
  • P4
  • M60
  • M10
  • M6

API Support on NVIDIA vGPU 14.0 

NVIDIA vGPU includes support for the following APIs:
  • Open Computing Language (OpenCL™ software) 3.0
  • OpenGL® 4.6
  • Vulkan® 1.3
  • DirectX 11
  • DirectX 12 (Windows 10)
  • Direct2D
  • DirectX Video Acceleration (DXVA)
  • NVIDIA® CUDA® 11.6
  • NVIDIA vGPU software SDK (remote graphics acceleration)
  • NVIDIA RTX (on GPUs based on the NVIDIA Volta graphic architecture and later architectures)
Note: These APIs are backwards compatible. Older versions of the API are also supported.

NVIDIA CUDA Toolkit and OpenCL Support on NVIDIA vGPU Software 14.0

NVIDIA CUDA Toolkit and OpenCL are supported with NVIDIA vGPU only on a subset of vGPU types and supported GPUs.

Note:

If you are using NVIDIA vGPU software with CUDA on Linux, avoid conflicting installation methods by installing CUDA from a distribution-independent runfile package. Do not install CUDA from a distribution-specific RPM or Deb package.

To ensure that the NVIDIA vGPU software graphics driver is not overwritten when CUDA is installed, deselect the CUDA driver when selecting the CUDA components to install.

OpenCL and CUDA Application Support

OpenCL and CUDA applications are supported on the following NVIDIA vGPU types:

  • The 8Q vGPU type on Tesla M6, Tesla M10, and Tesla M60 GPUs
  • All Q-series vGPU types on the following GPUs:
    • NVIDIA A2
    • NVIDIA A10
    • NVIDIA A16
    • NVIDIA A40
    • NVIDIA RTX A5000
    • NVIDIA RTX A6000
    • Tesla P4
    • Tesla P6
    • Tesla P40
    • Tesla P100 SXM2 16 GB
    • Tesla P100 PCIe 16 GB
    • Tesla P100 PCIe 12 GB
    • Tesla V100 SXM2
    • Tesla V100 SXM2 32GB
    • Tesla V100 PCIe
    • Tesla V100 PCIe 32GB
    • Tesla V100S PCIe 32GB
    • Tesla V100 FHHL
    • Tesla T4
    • Quadro RTX 6000
    • Quadro RTX 6000 passive
    • Quadro RTX 8000
    • Quadro RTX 8000 passive
  • All C-series vGPU types

NVIDIA CUDA Toolkit Development Tool Support

NVIDIA vGPU supports the following NVIDIA CUDA Toolkit development tools on some GPUs:

  • Debuggers:
    • CUDA-GDB
    • Compute Sanitizer
  • Profilers:
    • The Activity, Callback, and Profiling APIs of the CUDA Profiling Tools Interface (CUPTI)

      Other CUPTI APIs, such as the Event and Metric APIs, are not supported.

    • NVIDIA Nsight™ Compute
    • NVIDIA Nsight Systems
    • NVIDIA Nsight plugin
    • NVIDIA Nsight Visual Studio plugin

    Other CUDA profilers, such as nvprof and NVIDIA Visual Profiler, are not supported.

These tools are supported only in Linux guest VMs.

NVIDIA CUDA Toolkit profilers are supported and can be enabled on a VM for which unified memory is enabled.

Note:By default, NVIDIA CUDA Toolkit development tools are disabled on NVIDIA vGPU. If used, you must enable NVIDIA CUDA Toolkitdevelopment tools individually for each VM that requires them by setting vGPU plugin parameters. For instructions, see Enabling NVIDIA CUDA Toolkit Development Tools for NVIDIA vGPU.

The following table lists the GPUs on which NVIDIA vGPU supports these debuggers and profilers.

GPU vGPU Mode Debuggers Profilers
NVIDIA A2 Time-sliced ✓ ✓
NVIDIA A10 Time-sliced ✓ ✓
NVIDIA A16 Time-sliced ✓ ✓
NVIDIA A30 Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A30X Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A40 Time-sliced ✓ ✓
NVIDIA A100 HGX 40GB Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A100 PCIe 40GB Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A100 HGX 80GB Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A100 PCIe 80GB Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA A100X Time-sliced ✓ ✓
MIG-backed ✓ ✓
NVIDIA RTX A5000 Time-sliced ✓ ✓
NVIDIA RTX A6000 Time-sliced ✓ ✓
Tesla T4 Time-sliced ✓ ✓
Quadro RTX 6000 Time-sliced ✓ ✓
Quadro RTX 6000 passive Time-sliced ✓ ✓
Quadro RTX 8000 Time-sliced ✓ ✓
Quadro RTX 8000 passive Time-sliced ✓ ✓
Tesla V100 SXM2 Time-sliced ✓ ✓
Tesla V100 SXM2 32GB Time-sliced ✓ ✓
Tesla V100 PCIe Time-sliced ✓ ✓
Tesla V100 PCIe 32GB Time-sliced ✓ ✓
Tesla V100S PCIe 32GB Time-sliced ✓ ✓
Tesla V100 FHHL Time-sliced ✓ ✓

✓ Feature is supported

– Feature is not supported

Supported Hypervisors with NVIDIA vGPU 14.0

  • Citrix Hypervisor 7.1, 8.2 is supported with GPUs, M6, M10, M60, P4, P6, P40, P100, V100, RTX 6000, RTX 8000 Important: NVIDIA GPUs (A2, A10, A16, A30, A30X, A40, A100, A100X, T4,RTX A5000, RTX A6000, GPUs) are not supported with Citrix Hypervisor 7.1, only with Citrix Hypervisor 8.2.vGPU features: vCS, XenMotion with vGPU, Multiple vGPU on a single VM is not supported with Citrix Hypervisor 7.1 the mentioned features and GPU’s are only supported with Citrix Hypervisor 8.2.
  • VMware vSphere 6.5, 6.7, 7.0 is supported with GPUs, A2, A10, A16, A40, M6, M10, M60, P4, P6, P40, P100, V100, T4, V100, RTX A5000, RTX A6000, RTX6000, RTX8000, Note: Support for NVIDIA vGPU software requires the Enterprise Plus Edition of VMware vSphere Hypervisor (ESXi)Important: NVIDIA GPUs vGPU features are not supported with vSphere 6.5: Assignment of multiple vGPUs to a single VM Suspend-resume with vGPU vMotion with vGPU Live VMware snapshot with vGPU vSphere 6.7 U3, the assignment of multiple vGPUs to a single VM is supported. vSphere 6.7 supports only suspend and resume with vGPU. vMotion with vGPU is not supported on release 6.7.
  • Microsoft Hyper-V 2016 (1607, 1709, 1803) Microsoft Windows Server 2016 with Hyper-V role (supported with NVIDIA A2, A10, A16, A30, A30X, A40, A100, A100X, M6, M10, M60, P4, P6, P40, P100, V100, T4, V100, RTX A5000, RTX A6000, RTX6000, RTX8000) note: Microsoft Windows Server with Hyper-V role supports GPU pass-through over Microsoft Virtual PCI bus. This bus is supported through paravirtualized drivers.
  • Microsoft Server 2019 with Hyper-V role Microsoft Windows Server 2016 with Hyper-V role (supported with NVIDIA A2, A10, A16, A30, A30X, A40, A100, A100X, M6, M10, M60, P4, P6, P40, P100, V100, T4, V100, RTX A5000, RTX A6000, RTX6000, RTX8000) note: Microsoft Windows Server with Hyper-V role supports GPU pass-through over Microsoft Virtual PCI bus. This bus is supported through paravirtualized drivers.
  • Microsoft Server 2022 with Hyper-V role Microsoft Windows Server 2016 with Hyper-V role (supported with NVIDIA A2, A10, A16, A30, A30X, A40, A100, A100X, M6, M10, M60, P4, P6, P40, P100, V100, T4, V100, RTX A5000, RTX A6000, RTX6000, RTX8000) note: Microsoft Windows Server with Hyper-V role supports GPU pass-through over Microsoft Virtual PCI bus. This bus is supported through paravirtualized drivers.
  • Red Hat Enterprise Linux with KVM Red Hat Enterprise Linux with KVM 8.2, 8.4, 8.5 (supported with NVIDIA A2, A10, A16, A30, A30X, A40, A100, A100X, M6, M10, M60, P4, P6, P40, P100, V100, T4, RTX A5000, RTX A6000, RTX6000, RTX8000)
  • Red Hat Enterprise Linux with KVM 7.9 (supported with NVIDIA A2, A10, A16, A30, A30X, A40, A100, A100X, M6, M10, M60, P4, P6, P40, P100, V100, T4, RTX A5000, RTX A6000, RTX6000, RTX8000)

Supported Cloud Services for vGPU 14.0

NVIDIA virtual GPU software is supported on several cloud services with bring your own license (BYOL) licensing and licensing provided by the cloud service. Read more here: 
  • Alibaba Cloud
  • Amazon Web Services Elastic Compute Cloud (AWS EC2)
  • Google Cloud Platform (GCP)
  • Microsoft Azure

Alibaba Cloud

NVIDIA vGPU Support
GPU Supported Alibaba Cloud ECS Instances Supported Guest Operating Systems NVIDIA Licensing
Tesla P4 ecs.vgn5i-m1.large ecs.vgn5i-m2.xlarge ecs.vgn5i-m4.2xlarge ecs.vgn5i-m8.4xlarge Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows Server 2012 R2 Microsoft Windows Server 2008 R2 CentOS 7.x (64-bit) Red Hat Enterprise Linux 7.5 (64-bit) SUSE Linux Enterprise Server 11 SP4/12 SP2 (64-bit) Ubuntu 16.04 LTS (64-bit) BYOL
Tesla T4 ecs.vgn6i-m4.xlarge ecs.vgn6i-m8.2xlarge
GPU Pass Through Support
GPU Supported Alibaba Cloud ECS Instances Supported Guest Operating Systems NVIDIA Licensing
Tesla P4 ecs.gn5i-c2g1.large ecs.gn5i-c4g1.xlarge ecs.gn5i-c8g1.2xlarge ecs.gn5i-c16g1.4xlarge ecs.gn5i-c16g1.8xlarge ecs.gn5i-c28g1.14xlarge Windows Server 2019 Windows Server 2016 1607, 1709 Windows Server 2012 R2 Windows Server 2008 R2 CentOS 7.x (64-bit) Red Hat Enterprise Linux 7.4/7.5 (64-bit) SUSE Linux Enterprise Server 11 SP4/12 SP2 (64-bit) Ubuntu 16.04 LTS (64-bit) BYOL
Tesla P100 ecs.gn5-c4g1.xlarge ecs.gn5-c8g1.2xlarge ecs.gn5-c4g1.2xlarge ecs.gn5-c8g1.4xlarge ecs.gn5-c28g1.7xlarge ecs.gn5-c8g1.8xlarge ecs.gn5-c28g1.14xlarge ecs.gn5-c8g1.14xlarge
Tesla T4 ecs.gn6i-c4g1.xlarge ecs.gn6i-c8g1.2xlarge ecs.gn6i-c16g1.4xlarge ecs.gn6i-c24g1.6xlarge ecs.gn6i-c24g1.12xlarge ecs.gn6i-c24g1.24xlarge ecs.gn6i-c32g1.8xlarge ecs.gn6i-c48g1.12xlarge ecs.gn6i-c72g1.18xlarge
Tesla V100 ecs.gn6v-c8g1.2xlarge ecs.gn6v-c8g1.8xlarge ecs.gn6v-c8g1.16xlarge
 
Related Alibaba Cloud Documentation
Create a compute optimized instance with GPU vgn5i, light-weight compute optimized type family with GPU Install the GPU driver Install a GRID driver in a GPU-equipped ECS instance (Linux)

Amazon Web Services Elastic Compute Cloud (AWS EC2)

Only GPU pass through is supported on AWS EC2.
GPU Supported AWS EC2 Instances Supported Guest Operating Systems NVIDIA Licensing
Tesla M60 g3.4xlarge g3.8xlarge g3.16xlarge g3s.xlarge Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows Server 2012 R2 Microsoft Windows Server 2008 R2 CentOS 7.x (64-bit) Red Hat Enterprise Linux 7.5 (64-bit) SUSE Linux Enterprise Server 11 SP4/12 SP2 (64-bit) Ubuntu 16.04 LTS (64-bit) Provided by AWS
Tesla T4 g4dn.xlarge g4dn.2xlarge g4dn.4xlarge g4dn.8xlarge g4dn.12xlarge g4dn.16xlarge
Tesla A10G g5.xlarge g5.2xlarge g5.4xlarge g5.8xlarge g5.12xlarge g5.16xlarge g5.24xlarge g5.48xlarge
Tesla M60 g3.4xlarge g3.8xlarge g3.16xlarge g3s.xlarge Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows Server 2012 R2 Microsoft Windows Server 2008 R2 CentOS 7.x (64-bit) Red Hat Enterprise Linux 7.5 (64-bit) SUSE Linux Enterprise Server 11 SP4/12 SP2 (64-bit) Ubuntu 16.04 LTS (64-bit) BYOL
Tesla T4 g4dn.xlarge g4dn.2xlarge g4dn.4xlarge g4dn.8xlarge g4dn.12xlarge g4dn.16xlarge
Tesla V100 P3.2xlarge P3.8xlarge P3.16xlarge
Linux Accelerated Computing Instances Windows Accelerated Computing Instances

Google Cloud Platform (GCP)

GPU Supported GCP Instances Supported Guest Operating Systems NVIDIA Licensing
Tesla P4 Any predefined machine type Any custom machine type that can be created in a zone Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows Server 2012 R2 Red Hat Enterprise Linux 7.x (64-bit) Ubuntu 16.04 LTS/18.04 LTS (64-bit) Provided by GCP
Tesla P100
Tesla T4
Tesla P4 Any predefined machine type Any custom machine type that can be created in a zone Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows Server 2012 R2 Red Hat Enterprise Linux 7.x (64-bit) Ubuntu 16.04 LTS/18.04 LTS (64-bit) BYOL
Tesla P100
Tesla T4
Tesla V100
Related GCP Documentation
NVIDIA and Google Cloud Platform GPUs on Compute Engine

Microsoft Azure

GPU Pass Through Support
GPU Supported Microsoft Azure VM Sizes Supported Guest Operating Systems NVIDIA Licensing
Tesla M60 NV6 NV12 NV24 NV12s_v3 NV24s_v3 NV48s_v3 Microsoft Windows Server 2022 Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows 10 Microsoft Windows 11 Red Hat Enterprise Linux 7.x (64-bit) SUSE Linux Enterprise Server 12 SP2 Ubuntu 16.04 LTS/18.04 LTS (64-bit) Provided by Microsoft Azure
Tesla T4 NC4as_T4_v3 NC8as_T4_v3 NC16as_T4_v3 NC64as_T4_v3
Tesla M60 NV6 NV12 NV24 NV12s_v3 NV24s_v3 NV48s_v3 Microsoft Windows Server 2022 Microsoft Windows Server 2019 Microsoft Windows Server 2016 1607, 1709 Microsoft Windows 10 Microsoft Windows 11 Red Hat Enterprise Linux 7.x (64-bit) SUSE Linux Enterprise Server 12 SP2 Ubuntu 16.04 LTS/18.04 LTS (64-bit) BYOL
Tesla P40 ND6s ND12s ND24s
Tesla P100 NC6s_v2 NC12s_v2 NC24s_v2
Tesla V100 NC6s_v3 NC12s_v3 NC24s_v3
Related Microsoft Azure Documentation
GPU optimized virtual machine sizes NVIDIA GPU Driver Extension for Windows NVIDIA GPU Driver Extension for Linux Install NVIDIA GPU drivers on N-series VMs running Windows Install NVIDIA GPU drivers on N-series VMs running Linux
Related NVIDIA Knowledge Base Articles
Known issue: Microsoft Azure Linux image fails to acquire an NVIDIA virtual GPU software license

Supported Hypervisor with migration of vGPU 14.0 across hypervisor

  • XenMotion with vGPU is supported with Citrix XenServer 8.2
  • vMotion with vGPU is supported with VMware vSphere 6.7 CU2, CU3
Supported VMware vSphere Hypervisor (ESXi) releases:
  • Release 7.0 and compatible updates support vMotion with vGPU and suspend-resume with vGPU.
  • Release 6.7 U1 and compatible updates support vMotion with vGPU and suspend-resume with vGPU.
  • Release 6.7 supports only suspend-resume with vGPU.
  • Releases earlier than 6.7 do not support any form of vGPU migration.
Supported guest OS releases: Windows and Linux

Supported GPUs, that support vGPU Migration Support 

  • Tesla M6
  • Tesla M10
  • Tesla M60
  • Tesla P4
  • Tesla P6
  • Tesla P40
  • Tesla V100 SXM2
  • Tesla V100 SXM2 32GB
  • Tesla V100 PCIe
  • Tesla V100 PCIe 32GB
  • Tesla V100S PCIe 32GB
  • Tesla V100 FHHL
  • Tesla T4
  • Quadro RTX 6000
  • Quadro RTX 6000 passive
  • Quadro RTX 8000
  • Quadro RTX 8000 passive
  • NVIDIA A2
  • NVIDIA A10
  • NVIDIA A16
  • NVIDIA A40
  • NVIDIA RTX A5000
  • NVIDIA RTX A6000

What is MULTI-vGPU

Supported Hypervisor with Multiple vGPU support

Following Hypervisors is supported with assigning multiple vGPU to a single VM:

  • VMware vSphere 7.0, 6.7CU3 and later compatible updates only.
  • Citrix Hypervisor 8.2
  • RHEL KVM 7.9, 8.2, 8.3, 8.5
  • RHV 4.7
To support applications and workloads that are compute or graphics intensive, multiple vGPUs can be added to a single VM. The assignment of more than one vGPU to a VM is supported only on a subset of vGPUs

Supported vGPUs profile with (Multiple vGPU support functionality)

Only Q-series vGPUs that are allocated all of the physical GPU’s frame buffer are supported.

GPU Architecture Board vGPU
Ampere NVIDIA A40 A40-48Q See Note (1).
NVIDIA A16 A16-16Q See Note (1).
NVIDIA A10 A10-24Q See Note (1).
NVIDIA A2 A2-16Q See Note (1).
NVIDIA RTX A6000 A6000-48Q See Note (1).
NVIDIA RTX A5000 A5000-24Q See Note (1).
Turing Tesla T4 T4-16Q
Quadro RTX 6000 RTX6000-24Q
Quadro RTX 6000 passive RTX6000P-24Q
Quadro RTX 8000 RTX8000-48Q
Quadro RTX 8000 passive RTX8000P-48Q
Volta Tesla V100 SXM2 32GB V100DX-32Q
Tesla V100 PCIe 32GB V100D-32Q
Tesla V100S PCIe 32GB V100S-32Q
Tesla V100 SXM2 V100X-16Q
Tesla V100 PCIe V100-16Q
Tesla V100 FHHL V100L-16Q
Pascal Tesla P100 SXM2 P100X-16Q
Tesla P100 PCIe 16GB P100-16Q
Tesla P100 PCIe 12GB P100C-12Q
Tesla P40 P40-24Q
Tesla P6 P6-16Q
Tesla P4 P4-8Q
Maxwell Tesla M60 M60-8Q
Tesla M10 M10-8Q
Tesla M6 M6-8Q

Maximum vGPUs per VM

NVIDIA vGPU software supports up to a maximum of four vGPUs per VM on VMware vSphere Hypervisor (ESXi). VMware vSphere, Citrix Hypervisor, RHV or Red Hat Enterprise Linux with KVM.

Peer-to-Peer CUDA transfer over NVLink support

Peer-to-peer CUDA transfers enable device memory between vGPUs on different GPUs that are assigned to the same VM to be accessed from within the CUDA kernels. NVLink is a high-bandwidth interconnect that enables fast communication between such vGPUs. Peer-to-Peer CUDA Transfers over NVLink is supported only on a subset of vGPUs

Supported Hypervisor

Peer-to-Peer CUDA Transfers over NVLink are supported on all hypervisor releases that support the assignment of more than one vGPU to a VM

Supported Guest OS Releases

Linux only. Peer-to-Peer CUDA Transfers over NVLink are not supported on Windows.

Supported vGPUs

Only Q-series vGPUs that are allocated all of the physical GPU’s frame buffer on physical GPUs that support NVLink are supported.

GPU Architecture Board vGPU
Ampere NVIDIA A40 A40-48Q
NVIDIA A10 A10-24Q
NVIDIA RTX A6000 A6000-48Q
NVIDIA RTX A5000 A5000-24Q
Turing Quadro RTX 6000 RTX6000-24Q
Quadro RTX 6000 passive RTX6000P-24Q
Quadro RTX 8000 RTX8000-48Q
Quadro RTX 8000 passive RTX8000P-48Q
Volta Tesla V100 SXM2 32GB V100DX-32Q
Tesla V100 SXM2 V100X-16Q
Pascal Tesla P100 SXM2 P100X-16Q
 
Note:
  1. Supported only on the following hardware:
    • NVIDIA HGX™ A100 4-GPU baseboard with four fully connected GPUs

NVIDIA GPU Operator Support

NVIDIA GPU Operator simplifies the deployment of NVIDIA vGPU software with software container platforms on immutable operating systems. An immutable operating system does not allow the installation of the NVIDIA vGPU software graphics driver directly on the operating system. NVIDIA GPU Operator is supported only on specific combinations of VMware vSphere Hypervisor (ESXi) release, container platform, and guest OS release.

VMware vSphere Hypervisor (ESXi)Release Container Platform Guest OS
VMware vSphere Hypervisor (ESXi) 7.0 Update 2 VMware Tanzu Kubernetes Grid Ubuntu 20.04 LTS
VMware vSphere Hypervisor (ESXi) 7.0 Update 2 Red Hat Openshift 4.9 with Red Hat Enterprise Linux CoreOS and the CRI-Ocontainer runtime Red Hat CoreOS 4.9
VMware vSphere Hypervisor (ESXi) 7.0 Update 2 Red Hat Openshift 4.8 with Red Hat Enterprise Linux CoreOS and the CRI-Ocontainer runtime Red Hat CoreOS 4.8

NVIDIA Deep Learning Super Sampling (DLSS) Support

NVIDIA vGPU software supports NVIDIA DLSS on NVIDIA RTX Virtual Workstation.

Supported DLSS versions: 2.0. Version 1.0 is not supported.

Supported GPUs:
  • NVIDIA A40
  • NVIDIA A16
  • NVIDIA A2
  • NVIDIA A10
  • NVIDIA RTX A6000
  • NVIDIA RTX A5000
  • Tesla T4
  • Quadro RTX 8000
  • Quadro RTX 8000 passive
  • Quadro RTX 6000
  • Quadro RTX 6000 passive
Note: NVIDIA graphics driver components that DLSS requires are installed only if a supported GPU is detected during installation of the driver. Therefore, if the creation of VM templates includes driver installation, the template should be created from a VM that is configured with a supported GPU while the driver is being installed.

Supported applications: only applications that use nvngx_dlss.dll version 2.0.18 or newer

Whats new in NVIDIA Virtual GPU Manager 470.63

Included in this release are available for following hypervisors.
  • NVIDIA Virtual GPU Manager versions 470.63 for Citrix XenServer 8.2
  • NVIDIA Virtual GPU Manager versions 4470.63 for Citrix XenServer 7.1
  • NVIDIA Virtual GPU Manager version 470.63 for VMware vSphere 7.0 Hypervisor (ESXi)
  • NVIDIA Virtual GPU Manager version 470.63 for VMware vSphere 6.7 Hypervisor (ESXi)
  • NVIDIA Virtual GPU Manager version 470.63for VMware vSphere 6.5 Hypervisor (ESXi)
  • NVIDIA Virtual GPU Manager version 470.63 for Red Hat Enterprise Linux with KVM 7.7, 7.8, 7.9, 8.1, 8.2, 8.3, 8.4
  • NVIDIA Virtual GPU Manager version 470.63 for Red Hat Virtualization RHV 4.2, 4.3, 4.4

Windows Guest OS support in NVIDIA vGPU 14.0 – 471.68

NVIDIA vGPU 471.68 supports following Windows release as a guest OS
  • Windows Server 2022
  • Windows Server 2019
  • Windows Server 2016 1607, 1709
  • Windows 10 November Update (21H2) and all Windows 10 releases supported by Microsoft up to and including this release

Linux Guest OS support in NVIDIA vGPU 14.0 – 470.63.01

NVIDIA vGPU 470.63.01 supports following Linux distributions as a guest OS only on supported Tesla GPUs
  • Red Hat Enterprise Linux 7.9, 8.2, 8.4, 8.5
  • CentOS 8.0, 8 (1911)
  • Ubuntu 20.04 LTS
  • Ubuntu 18.04 LTS
  • Ubuntu 16.04 LTS
  • Ubuntu 14.04 LTS
  • Debian 10
  • CentOS 6.6
  • Red Hat CoreOS 4.7
  • SUSE Linux Enterprise Server 15 SP2
  • SUSE Linux Enterprise Server 12 SP2

Source

https://docs.nvidia.com https://www.nvidia.com/en-us/drivers/vgpu-software-driver https://docs.nvidia.com/grid/14.0/

Leave a Reply

Your email address will not be published. Required fields are marked *

Turn on pictures to see the captcha *