wikiNVIDIA vGPU Licensing
For the use of NVIDIA vGPU software for the virtualisation of GPU resources, different licenses are required depending on the use case. This article explains different licenses and explains the functions that can be used depending on the license.
Licenses
In the following, the licensing for different application scenarios is explained.
NVIDIA vApps
The vApps software is for all session based solutions and app streaming especially Citrix Virtual Apps or RDSH (Remote Desktop Session Hosts). The client is therefore only offered individual applications during its current session and not a complete remote desktop.
NVIDIA vPC
The vPC software is used if the clients are to be provided with complete Windows desktops including browser, media usage and other Windows applications.
NVIDIA RTX vWS
The RTX vWS software is used when computationally intensive workloads, for example graphics acceleration, are used. Possible applications are SOLIDWORKS, 3DExcite, Siemens NX and other 3D development tools.
Functions
The licenses listed above each contain the following tabulated functions[1]:
| feature | vApps | vPC | vWS |
|---|---|---|---|
| license permissions | |||
| concurrent user | ✔ | ✔ | ✔ |
| per GPU | ✘ | ✘ | ✘ |
| functions included | |||
| desktop virtualisation | ✘ | ✔ | ✔ |
| RDSH app hosting | ✔ | ✔ | ✔2 |
| RDSH desktop hosting | ✔ | ✔ | ✔2 |
| compute virtualisation | ✘ | ✘ | ✔ |
| Windows guest OS support | ✔ | ✔ | ✔ |
| Linux guest OS support | ✔ | ✔ | ✔ |
| maximum number of displays | 13 | 4 | 4 |
| maximum resolution4 | 1280 x 1024 | 5120 x 2880 (5K) | 7680 x 4320 (8K) |
| NVIDIA RTX enterprise
software features |
✘ | ✘ | ✔ |
| OpenGL, DirectX and Vulkan | ✔ | ✔ | ✔ |
| CUDA and OpenCL support | ✘ | ✘ | ✔5 |
| ECC and Page retirement | ✘ | ✘ | ✔ |
| Multi-vGPU | ✘ | ✘ | ✔ |
| NVLink | ✘ | ✘ | ✔ |
| GPU pass through support7 | ✔ | ✘ | ✔ |
| Bare Metal support8 | ✔ | ✘ | ✔ |
| supported vGPU profile sizes9 | |||
| 512 MB | ✘ | ✔ | ✔ |
| 1 GB | ✔ | ✔ | ✔ |
| 2 GB | ✔ | ✔ | ✔ |
| 3 GB | ✔ | ✘ | ✔ |
| 4 GB | ✔ | ✘ | ✔ |
| 5 GB | ✘ | ✘ | ✘ |
| 6 GB | ✔ | ✘ | ✔ |
| 8 GB | ✔ | ✘ | ✔ |
| 10 GB | ✘ | ✘ | ✘ |
| 12 GB | ✔ | ✘ | ✔ |
| 16 GB | ✔ | ✘ | ✔ |
| 20 GB | ✘ | ✘ | ✘ |
| 24 GB | ✔ | ✘ | ✔ |
| 32 GB | ✔ | ✘ | ✔ |
| 40 GB | ✘ | ✘ | ✘ |
| 48 GB | ✔ | ✘ | ✔ |
hints
1maximum 10 users at the same time per GPU
2with packaged NVIDIA vApps license
3only for the console display in remote application environments. Details can be found at supported GPUs
4Please inform yourself at Virtual GPU Software User Guide about the supported display configurations for every profile
5supported on 8GB 1:1-Profil on NVIDIA Maxwell and all profiles on Pascal
6ECC support starts with Pascal
7support only for 1:1 profiles
8only NVIDIA M6-hardware is supported as primary display device
9Please inform yourself at Virtual GPU Software User Guide about the graphics processors supported by your vGPU-Profile
License models
All software licenses are licensed per concurrent user (CCU - Concurrent Users). For licensing, either an annual subscription must be taken out, which must be renewed every year, or a perpetual license must be purchased.
Subscription
If you purchase the usage rights as a subscription, you can take out subscriptions for 1, 4 or 5 years for vPC and vWS licenses. The vApps licenses must be renewed annually.
All license costs are calculated per CCU.
Furthermore, there is a SUMS (Support, Upgrade and Maintenance Service) included in every license.
Perpetual
If you purchase permanent usage rights, these are also charged per CCU.
In contrast to the subscription model, SUMS must also be booked as an annual subscription with this variant.
Education price model
For educational and research institutions, NVIDIA offers a separate pricing structure for the NVIDIA vWS software.
There is also a subscription and perpetual model available[1].
Application cases
The most common application cases for the respective licenses are listed below.
The chart shows, which licenses are required depending on the use of the application:
| application | vApps | vPC | vWS |
|---|---|---|---|
| Citrix Virtual desktops | ✘ | ✔ | ✔ |
| Omnissa Horizon 8 | ✘ | ✔ | ✔1 |
| Citrix Virtual apps and desktops | ✔ | ✘ | ✔1 |
| Omnissa Horizon 8 RDSH | ✔ | ✘ | ✔1 |
| other session based / RDSH application | ✔ | ✘ | ✔1 |
| Microsoft RemoteFX | ✘ | ✔ | ✘ |
| VMware configurevSGA | ✘ | ✔ | ✘ |
| Microsoft Hyper-V (DDA) | ✘ | ✘ | ✔ |
| Microsoft AzureStack GPU-P | ✘ | ✔ | ✔1 |
| RedHat enterprise Linux with KVM | ✘ | ✘ | ✔ |
| Proxmox VE with KVM | ✘ | ✘ | ✔2 |
hint
1applies to every session that uses a workstation or a professional 3D application
2Proxmox VE is not officially supported by NVIDIA.
More information
- Grid vGPU User guide (nvidia.com, november 2024)
- Virtual GPU Software Client Licensing (nvidia.com, october 2024)
References
- ↑ 1.0 1.1 Virtual GPU Packaging and Licensing Guide (nvidia.com, November 2024)
 |
Author: Stefan Bohn Stefan Bohn has been employed at Thomas-Krenn.AG since 2020. Originally based in PreSales as a consultant for IT solutions, he moved to Product Management in 2022. There he dedicates himself to knowledge transfer and also drives the Thomas-Krenn Wiki.
|
