Method and device for realizing graphics processor virtualization

Abstract

The embodiment of the present invention discloses a method and device for realizing graphics processor virtualization, wherein the method includes initiating multiple parallel application execution instructions in response to a service request initiated by a remote concurrent graphics application; using the GPU hardware driver layer to provide call interface for hardware resources, and divide the GPU hardware resources in the cloud according to the requirements of concurrent applications, so as to allocate independent hardware resources for each application; run multiple parallel applications based on the allocated hardware resource environment, and complete the graphics of each Rendering: Know the hardware resource address used by each application through the path addressing mechanism, and directly intercept the rendered output data at the bottom of the hardware based on the known address; stream the rendered output data and transmit it to the initiator The remote user of each concurrent graphics application can directly display the graphics corresponding to each graphics application on the terminal used by the user. The embodiments of the present invention can provide high-end graphic applications for weak terminals.

Description

technical field [0001] The present invention relates to the field of cloud computing basic technology virtualization, in particular to a method and device for realizing graphics processor virtualization. Background technique [0002] With the continuous improvement of GPU (Graphics Processing Unit, graphics processing unit) performance, it is no longer limited to graphics processing. Its general computing capabilities have more prominent features in floating point operations and parallel computing. CPU (Central Processing Unit, central processing unit), which is not yet complete in terms of application virtualization, is difficult to effectively support the field of cloud computing. [0003] Compared with CPU, GPU has advantages in computing unit in terms of hardware architecture, but compared with mature CPU technology in terms of virtualization driver support and multi-core design, GPU itself has a large gap, and it is difficult to exert its powerful graphics rendering abi...

AI Technical Summary

Problems solved by technology

The main obstacle is that the existing terminal set-top boxes in the market are not capable enough to support high value-added services

The use of cloud computing technology can well solve the defect of weak user terminal capabilities, but at the same time it brings challenges to cloud implementation technology. Therefore, there is an urgent need for intelligent GPU virtual solutions suitable for graphics data processing to support business development.

Images