https://github.com/open-edge-platform/edge-desktop-virtualization
Edge-Desktop-Virtualization has essential components for the IDV use case, leveraging Intel's graphics and display virtualization technology (Graphics SR-IOV) on cloud/edge-native infrastructure (k3s/k8s/kubevirt)).
https://github.com/open-edge-platform/edge-desktop-virtualization
gfx-sriov intelligent-desktop-virtualization k3s kubernetes kubevirt linux-vm qemu-kvm sriov virtualization vm windows-vm
Last synced: 2 months ago
JSON representation
Edge-Desktop-Virtualization has essential components for the IDV use case, leveraging Intel's graphics and display virtualization technology (Graphics SR-IOV) on cloud/edge-native infrastructure (k3s/k8s/kubevirt)).
- Host: GitHub
- URL: https://github.com/open-edge-platform/edge-desktop-virtualization
- Owner: open-edge-platform
- Created: 2025-05-23T07:28:54.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2025-07-31T10:07:13.000Z (2 months ago)
- Last Synced: 2025-07-31T13:57:12.504Z (2 months ago)
- Topics: gfx-sriov, intelligent-desktop-virtualization, k3s, kubernetes, kubevirt, linux-vm, qemu-kvm, sriov, virtualization, vm, windows-vm
- Language: Shell
- Homepage:
- Size: 58.2 MB
- Stars: 1
- Watchers: 1
- Forks: 7
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSES/Apache-2.0.txt
- Code of conduct: CODE_OF_CONDUCT.md
- Security: SECURITY.md
Awesome Lists containing this project
README
# Desktop Virtualization with Graphics SR-IOV
[](https://opensource.org/licenses/Apache-2.0)
[](https://scorecard.dev/viewer/?uri=github.com/open-edge-platform/edge-desktop-virtualization)
[](https://github.com/open-edge-platform/edge-desktop-virtualization/actions/workflows/github-code-scanning/codeql)
[](https://github.com/open-edge-platform/edge-desktop-virtualization/actions/workflows/device_plugin_coverity.yaml)
- [Desktop Virtualization with Graphics SR-IOV](#desktop-virtualization-with-graphics-sr-iov)
- [Overview](#overview)
- [How it works](#how-it-works)
- [Key Features](#key-features)
- [System requirements:](#system-requirements)
- [Recommended Hardware Configuration](#recommended-hardware-configuration)
- [Host OS Options](#host-os-options)
- [Generate Virtual Machine qcow2 with required drivers for SR-IOV](#generate-virtual-machine-qcow2-with-required-drivers-for-sr-iov)
- [K3s/Kubevirt Solution stack for Edge Desktop Virtualization](#k3skubevirt-solution-stack-for-edge-desktop-virtualization)
- [IDV Services](#idv-services)
- [Device Plugins for Kubernetes](#device-plugins-for-kubernetes)
- [Kubevirt Patch](#kubevirt-patch)
- [Sample Application : VM deployment Helm charts](#sample-application--vm-deployment-helm-charts)
- [References](#references)
- [Getting Help](#getting-help)
- [Contributions](#contributions)
- [License](#license)
## Overview
Intel's Single Root I/O Virtualization (SR-IOV) for graphics is a technology that allows a single physical Intel graphics processing unit (GPU) to be presented as multiple virtual devices to different virtual machines (VMs). This enables efficient GPU resource sharing and improves performance for graphics-intensive workloads within virtualized environments
### How it works
- The Physical Function (PF) is the full-featured physical function of the GPU, managed by the host OS.
- Virtual functions (VFs) are created from the PF and assigned to individual VMs.
- Each VF provides a dedicated and isolated path for data transfer to and from the VM, bypassing the host's hypervisor for improved performance and reducded latency.
- PCIe Standard: SR-IOV is a PCI-SIG standard that defines how a single physical PCIe device can be presented as multiple virtual devices.
Graphics SR-IOV Architecture
### Key Features
- **Improved performance:** Direct access to the GPU hardware for each VM reduces overhead and latency, particularly for tasks like video transcoding and media processing.
- **Efficient resource utilization:** SR-IOV enables better sharing of GPU resources among multiple VMs, maximizing the utilization of a single physical GPU.
- **Support for cloud-native environments:** SR-IOV is crucial for enabling GPU acceleration in Kubernetes and other cloud platforms.
### System requirements:
SR-IOV for Intel graphics typically requires hardware generation of Alder Lake (12th Gen Intel Core) or newer.
One can check if your Intel graphics controller supports SR-IOV by executing below command for the Single Root I/O Virtualization (SR-IOV) PCI capability
```sh
sudo lspci -s 2.0 -v
```
#### Recommended Hardware Configuration
| Component | Edge Microvisor Toolkit + graphics SR-IOV|
|--------------|------------------------------------------|
| CPU | Intel® Core (12th gen and higher) |
| RAM | 64GB recommended |
| Storage | 500 GB SSD or NVMe minimum |
| Networking | 1GbE Ethernet |
### Host OS Options
Host OS has to be enabled with graphics SR-IOV ingredients.
Below are various options (but not limited to) :
- #### EMT - Intel EMT with [desktop-virtualization Image config](https://github.com/open-edge-platform/edge-microvisor-toolkit/blob/70d364596e88ad332637d7073a7a0a445960ca39/toolkit/imageconfigs/edge-image-desktop-virtualization.json)
- #### Debian - [ThunderSoft SR-IOV](https://github.com/ThunderSoft-SRIOV)
- #### Ubuntu
- [kvm multi-os](https://github.com/intel/kvm-multios/blob/main/documentation/setup_sriov.md)
- [12th Gen](https://www.intel.com/content/www/us/en/secure/content-details/680834/12th-gen-intel-core-mobile-processors-code-named-alder-lake-p-12th-gen-intel-core-desktop-processors-code-named-alder-lake-s-multi-os-with-graphics-sr-iov-virtualization-on-ubuntu-user-guide.html?wapkw=multi-os%20graphics%20SRIOV&DocID=680834)
- [13th Gen](https://www.intel.com/content/www/us/en/secure/content-details/762237/13th-gen-intel-core-mobile-processors-for-iot-edge-code-named-raptor-lake-p-multi-os-with-graphics-sr-iov-virtualization-on-ubuntu-user-guide.html?wapkw=multi-os%20graphics%20SRIOV)
- [Core Ultra PS Series (Meteor Lake)](https://www.intel.com/content/www/us/en/secure/content-details/780205/reference-implementation-of-intel-core-ultra-processor-intel-core-ultra-processor-ps-series-formerly-known-as-meteor-lake-u-h-ps-multi-os-with-graphics-sr-iov-virtualization-on-ubuntu-user-guide.html?wapkw=multi-os%20graphics%20SRIOV)
### Generate Virtual Machine qcow2 with required drivers for SR-IOV
- #### [Windows](https://github.com/ThunderSoft-SRIOV/sriov/blob/main/docs/deploy-windows-vm.md#microsoft-windows-11-vm)
- #### [Ubuntu](https://github.com/ThunderSoft-SRIOV/sriov/blob/main/docs/deploy-ubuntu-vm.md)
## K3s/Kubevirt Solution stack for Edge Desktop Virtualization
Below are the ingredients to achieve Display and Graphics Virtualization pipeline using SR-IOV.
Desktop Virtualization on K3s/Kubevirt
The components marked in red are in scope of this solution.
The display and graphics virtualization pipeline on k3s/kubevirt can be realized by following build and installation of below components.
### [IDV Services](idv-services/README.md)
### [Device Plugins for Kubernetes](device-plugins-for-kubernetes/README.md)
### [Kubevirt Patch](kubevirt-patch/README.md)
### Sample Application : VM deployment Helm charts
- #### [Discrete Helm charts](sample-application/discrete/README.md)
- #### [Single Helm deployment](sample-application/single/README.md)
## References
- [Reference-1](https://cyberus-technology.de/en/articles/vbox-kvm-sriov)
- [Reference-2](https://www.dfi.com/Uploads/DownloadCenter/5631e304-28b2-4256-975a-5689750b5636/Intel%20iGPU%20(Integrated%20Graphics)%20SR-IOV%20-%20The%20Catalyst%20for%20IoT%20Virtualization%20in%20Factory%20Automation.pdf?timestamp=1676441838.9072)
- [Reference-3](https://www.intel.com/content/www/us/en/secure/content-details/782115/intelligent-desktop-virtualization-idv.html?wapkw=intelligent%20desktop%20virtualization&DocID=782115)
- [Reference-4](https://www.intel.com/content/www/us/en/secure/content-details/657261/sr-iov-based-graphics-virtualization.html?wapkw=intelligent%20desktop%20virtualization&DocID=657261)
- [References-5](https://github.com/intel/kubevirt-gfx-sriov)
- [References-6](https://dgpu-docs.intel.com/devices/hardware-table.html)
## Getting Help
if you encounter bugs, have feature requests, or need assistance, [file a Github issue](https://github.com/open-edge-platform/edge-desktop-virtualization/issues)
For security-related concerns, please refer to [SECURITY.md](./SECURITY.md).
## Contributions
Edge Desktop Virtualization is open source and always welcomes an active community to support adding new features, optimizing, and improving security.
To learn how to contribute to the project, see the [Contributor's Guide](./CONTRIBUTING.md).
## License
For license related information refer [LICENSES](./LICENSES/)
Last Updated Date: June 02, 2025