Selecting the Correct vGPU Profiles
Choosing the right profile to maximize your stakeholder’s experience within the virtual instance. Below, you will find guidance through the vGPU Manager and beyond to ensure your successful deployment.
NVIDIA vGPU profiles assign custom amounts of dedicated graphics memory to each virtual machine. NVIDIA vGPU Manager assigns the correct amount of memory to meet the specific needs within the workflow for the virtual machine user. Every virtual machine has dedicated graphics memory and must be assigned accordingly, thus ensuring that it has the resources needed to handle the expected graphics load.
NVIDIA vGPU Manager allows multiple users to share each physical GPU by assigning the graphics resources of the available GPUs to virtual machines using a balanced approach. Depending on the number of GPUs within each NVIDIA card, multiple user types can be assigned.
vGPU profiles represent very flexible deployment options for virtual GPUs, varying the allocated frame buffer memory size depending on several factors, including the number and resolution of display heads. The division of the frame buffer defines the number of users possible per GPU with that specific profile, while the number of heads establishes the number of displays supported. Max resolution is consistent across all the profiles. The complete list may also be found here.
Series | Optimal Workload |
|
|---|---|
| Q-series | Virtual workstations for creative and technical professionals who require the performance and features of RTX Enterprise Drivers. |
| B-series | Virtual desktops for business professionals and knowledge workers |
| A-series | App streaming or session-based solutions for virtual applications users |
NVIDIA vGPU is a licensed product on all supported GPU boards. A software license is required to enable all vGPU features within the guest VM, and the type of license required depends on the vGPU type.
Q-series vGPU types require an RTX vWS license.
B-series vGPU types require an NVIDIA Virtual PC license but can also be used with an RTX vWS license.
A-series vGPU types require a NVIDIA Virtual Applications license
Starting with vGPU 17, NVIDIA vGPU software supports heterogeneous configurations, allowing different types of time-sliced vGPUs to be used simultaneously on the same physical GPU. This means that a mixture of A-series, B-series, and Q-series vGPUs with varying amounts of frame buffer can coexist on the same GPU, provided the total frame buffer allocated does not exceed the physical GPU’s capacity.
By default, a GPU operates in equal-size mode, where all vGPUs on the GPU must have the same amount of frame buffer. To enable vGPUs with different frame buffer sizes to share a single GPU, the GPU must be switched to mixed-size mode. In mixed-size mode, the GPU can host vGPUs with varying frame buffer sizes, but the maximum number of some types of vGPUs allowed on the GPU may be reduced compared to equal-size mode.
For more details on enabling mixed-size mode and supported configurations, refer to:
Mixed-size mode is supported on VMware vSphere but not on XenServer. To determine whether your chosen hypervisor supports this feature with your specific GPU, consult the release notes for your hypervisor and the NVIDIA Virtual GPU Software Documentation.
A NVIDIA Virtual Application license is required for Citrix Virtual Application deployments. A NVIDIA Virtual PC license is required for Citrix Virtual Desktop deployments.