Shared Memory for GPU Virtualization Data Exchange
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In GPU virtualization, the transmission of graphics processing instructions and data between a virtual desktop and a GPU virtual rendering server results in long delays, low real-time response speed, and increased bandwidth due to the need for network transmission, affecting graphics processing performance.
Innovation Solution
A method where a virtual client process and a virtual rendering server create a shared memory, allowing the virtual rendering server to access graphics instruction data and send target display data through this shared memory, reducing the need for direct network transmission by using storage address information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If graphics processing instructions and data are transmitted through a network transmission channel between virtual desktop and GPU virtual rendering server, then data exchange is enabled, but transmission delay increases and real-time response speed decreases
Solution Approach 1:
The patent merges the virtual desktop's memory space with the GPU virtual rendering server's memory space by creating a shared memory area. This allows both components to access the same memory region directly, eliminating the need for network transmission of graphics processing instructions and data, thus reducing transmission delay while maintaining data exchange capability.
Solution Approach 2:
The patent introduces a shared memory area as an intermediary between the virtual desktop and GPU virtual rendering server. This shared memory serves as a common access space that both components can read from and write to directly, replacing the network transmission channel and enabling fast data exchange without network protocol overhead.
2Reliability
If all graphics processing instructions and rendering data are transmitted through network channel, then complete data exchange is achieved, but bandwidth consumption increases
Solution Approach 1:
The patent combines the memory resources of the virtual desktop and GPU virtual rendering server into a shared memory space. This merging eliminates the need for continuous network transmission of graphics data, thereby reducing bandwidth consumption while ensuring complete data exchange through direct memory access.
3Adaptability or versatility
If network transmission is used for graphics data exchange, then remote access is enabled, but graphics processing performance deteriorates
Solution Approach 1:
The patent merges memory spaces to enable direct access between virtual desktop and GPU rendering server, eliminating network transmission bottlenecks. This significantly improves graphics processing performance by allowing fast data exchange, while the virtualization framework preserves remote access capability through the virtual desktop interface.
Data Source
Figure 1~2
Figure 3~4
Figure 5A
AI summary
Embodiments of the present invention disclose a graphics data processing method, apparatus, and system. In the embodiments of the present invention, a virtual client process and a virtual rendering server may create a shared memory through negotiation. The virtual client process may send storage address information of graphics instruction data to the virtual rendering server, so that the virtual rendering server may obtain the graphics instruction data from the shared memory according to the storage address information of the graphics instruction data. The virtual rendering server may further send storage address information of target display data to the virtual client process, so that the virtual client process may obtain the target display data from the shared memory according to the storage address information of the target display data. In this way, the graphics instruction data or the target display data or both do not need to be directly transmitted through a network transmission channel between the virtual client process and the virtual rendering server. This shortens a transmission delay, decreases bandwidth of the transmission channel, and improves graphics processing performance.