Virtual GPU Remote 3D Rendering Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern automotive vehicles require multiple computers to display 3D applications due to a lack of physical connectors, leading to inefficiencies in existing remote display sharing solutions that consume high bandwidth and suffer from poor 3D performance, as they often require GPU rendering and lack support for remote control units.
Innovation Solution
A system comprising a first control unit that transmits OpenGL commands over a network to a second control unit, utilizing a virtual GPU device to transform and send graphic commands in sequential data, reducing the need for extensive bandwidth and enabling efficient remote rendering of 3D applications across multiple displays within a vehicle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If OpenGL commands are sent over the network using GLX, then 3D applications can be displayed on remote displays, but the system consumes high bandwidth and is very slow
Solution Approach 1:
The patent extracts the rendering function from the remote display system and concentrates it in a local rendering node. Instead of sending rendering commands over the network, the system sends only the final rendered image data, separating the computationally intensive rendering process from the display process. This resolves the contradiction by maintaining remote display capability while dramatically improving speed.
Solution Approach 2:
The patent inverts the traditional remote display architecture by placing the GPU and rendering capabilities at the display node rather than the source node. The display system includes a local GPU that can independently render graphics, receiving only minimal control data from the source computer. This inversion eliminates the bandwidth and speed problems of traditional approaches.
2Productivity
If remote display solutions use GPU rendering at the server, then 3D performance can be improved, but the system requires complex GPU resource management and lacks support for remote control units
Solution Approach 1:
The patent segments the system into distinct functional nodes: a source computer that generates content, a network transmission layer, and display nodes with integrated GPUs. Each node performs its specialized function independently, eliminating the need for complex centralized GPU management while maintaining high 3D performance at each display location.
3Adaptability or versatility
If multiple computers are used to display 3D applications, then physical connector limitations are overcome, but network bandwidth consumption increases and latency is introduced
Solution Approach 1:
The patent applies preliminary action by rendering the complete image data before transmission. The source computer renders the entire frame in advance, compresses it, and sends it as a complete unit over the network. This eliminates the need for continuous incremental updates and reduces the impact of network latency, as the display node receives ready-to-display data.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a method for operating a first control unit (1, 1a, 1b, 1a', 1b', 42) in order to transmit display commands over a computer network (13) to a second control unit (1, 1a, 1b, 1a', 1b', 46), comprising: generating, by an application (20a, 20b) at the first control unit, at least one display command, the at least one display command enabling a rendering of a frame comprising a three-dimensional image by a graphic processing unit (7a, 7b); receiving, by a graphic processing unit driver (22a, 22b) on the first control unit, the at least one display command; providing, by the graphic processing unit driver (22a, 22b) the at least one display command or graphic commands being transformed from the display commands to a virtual graphic processing unit device (26a, 26b); collecting, by the virtual graphic processing unit device (26a, 26b), a plurality of display commands or graphic commands being transformed from the display commands; determining parameters of the collected plurality of display commands or graphic commands stored in at least one section of the at least one memory (5) using at least one reference to the at least section of the at least one memory (5); aligning the display commands or graphic commands and their in sequential data, wherein the references to the at least one section of the at least one memory (5) is transformed to an offset within the sequential data; encapsulating, by virtual graphic processing unit device, the created sequential data in one or more network packets; transmitting the at least one network packet by the virtual graphic processing unit device (26a, 26b) over the computer network to a second control unit adapted to generate display frames of at least one image according to the display commands.