What is a vGPU? How Virtual GPUs Work
A vGPU lets multiple users or virtual machines share a physical GPU without giving each user access to the entire card. It is useful when several workloads need GPU acceleration but dedicating one physical GPU to every user would be inefficient.
What Is a vGPU?
A virtual GPU(vGPU), is a portion of a physical GPU assigned to a virtual machine or user. The physical GPU is partitioned into dedicated slices, and each user gets their own isolated VRAM and GPU resources.
For example, one physical GPU can provide several vGPUs. Each virtual machine sees its assigned GPU resources rather than the full physical card. This allows multiple users to use the same GPU at the same time.
A vGPU works differently from simply sharing a GPU between applications. The GPU is divided into separate resources that can be assigned to individual virtual machines.
How Does vGPU Work?
A physical GPU is installed in a host system. Virtualization software and supported GPU technology divide its resources into multiple virtual GPUs.
- Physical GPU: The host contains the actual GPU hardware.
- GPU partitioning: The physical GPU is divided into multiple dedicated slices.
- Virtual machines: Each VM receives an assigned vGPU.
- Dedicated VRAM: Each vGPU has its own allocated VRAM.
- Isolation: Users operate within their assigned GPU resources rather than accessing another user's vGPU.
The exact number and size of vGPUs available depends on the physical GPU and the virtualization technology being used.
vGPU vs a Dedicated GPU
| Features | Dedicated GPU | vGPU |
|---|---|---|
| GPU allocation | One user or VM gets the physical GPU. | Multiple users or VMs share one physical GPU through separate vGPUs. |
| VRAM | The user can access the GPU's available VRAM. | Each vGPU receives its own allocated VRAM. |
| Users per GPU | Typically one. | Multiple, depending on the GPU and configuration. |
| Best suited for | Workloads that need substantial GPU resources. | Multiple workloads that need dedicated portions of a GPU. |
A dedicated GPU makes more sense when a workload needs most or all of the card's resources. vGPU is useful when several users need GPU acceleration but do not each need an entire physical GPU.
What Can You Use a vGPU For?
vGPUs can support many workloads that benefit from GPU acceleration. The appropriate vGPU size depends on the software and workload.
- AI and machine learning workloads
- 3D applications and engineering software
- Video editing
- Software development that uses GPU acceleration
- Remote workstations
- Cybersecurity and other technical workloads
For larger AI models, video projects, or demanding 3D applications, the amount of available VRAM can be an important factor when choosing a GPU or vGPU configuration.
Why Use vGPUs in Cloud Desktops?
Cloud desktops can use vGPUs to provide GPU-accelerated virtual machines to multiple users from the same physical hardware. This can make better use of a GPU when individual users do not need an entire card.
For example, a team could use separate virtual desktops while sharing the resources of a physical GPU through dedicated vGPU allocations. Each user gets their own virtual GPU and isolated VRAM instead of sharing a single desktop environment.
Try on DaDesktop
DaDesktop provides cloud desktops with dedicated GPUs and vGPU options for workloads that need GPU acceleration. Learn more about DaDesktop cloud GPU desktops.