A cloud desktop display card sharing method and system based on a GPU container technology

Abstract

The invention relates to a cloud desktop display card sharing method and system based on a GPU (Graphics Processing Unit) container, which are used for solving the technical problems of fixed video memory capacity, high use cost and poor compatibility caused by an existing method for realizing display card sharing of a plurality of cloud desktops based on a display card virtualization technology.According to the invention, the physical display card is abstracted into the GPU container and is bound to the cloud host; According to the method, a plurality of cloud desktops under a cloud host canshare a GPU container bound with the cloud host so as to realize the sharing of virtual graphics cards, common graphics cards or game graphics cards of the plurality of cloud desktops, and video memory resources can be dynamically allocated to each cloud desktop according to the self load of each cloud desktop; According to the invention, the idle condition of the display card resources can be avoided, the display card resource utilization rate is improved, and the use cost is reduced.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a method and system for sharing graphics cards on cloud desktops based on GPU container technology. Background technique [0002] Due to the advantages of simple management, convenient use, and high cost performance, cloud desktops have gradually replaced the use of traditional personal computers. In common office use scenarios, cloud desktops distribute CPU, memory and other resources of a physical host to multiple cloud desktops for sharing. It can be used by multiple users. In scenarios with special requirements for graphic images and video display, in addition to CPU and memory resources, the cloud desktop also needs to share graphics card resources with multiple users. [0003] In the existing technology, graphics card sharing in cloud desktops is usually realized based on graphics card virtualization technology (vGPU technology), which virtualizes a physical graph...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, graphics card sharing in cloud desktops is usually based on graphics card virtualization technology (vGPU technology), which virtualizes one physical graphics card device into multiple Virtual graphics card device, and then assign each virtual graphics card device to the cloud desktop that needs to use the graphics card, one virtual graphics card corresponds to one cloud desktop; this method can realize the function of multiple cloud desktops sharing a physical graphics card, but it has three There are two main problems: (1) After the virtual graphics card is set up, the video memory resources of the graphics card are fixed, and the corresponding cloud desktop cannot dynamically use larger or smaller video memory resources; (2) The compatibility is not good, only supports the market There are a few limited virtual graphics cards in the market, which do not support ordinary graphics cards or game graphics cards; (3) The cost of use is high, and virtual graphics cards are more expensive than graphics cards or game graphics cards

Images