Hybrid Graphics Display Data Routing via IGPU Raster Generator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for utilizing multiple graphics processing units (GPUs) in computers face issues such as power consumption, display disruptions, and performance bottlenecks when switching between integrated GPUs (IGPUs) and discrete GPUs (dGPUs), particularly due to data duplication and bandwidth concerns across the PCI-E bus.
Innovation Solution
The approach involves transmitting scan-out data from one graphics subsystem to the raster generator of another for display, allowing graphics data rendered by a dGPU to be output to a monitor connected to an IGPU, enabling multiple displays to be utilized without regard to which GPU they are connected to, and eliminating the need for backend switching and data duplication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a backend switching mechanism is used to switch between IGPU and dGPU, then the system can select which GPU drives the display, but the user experiences display disruption and the switching mechanism adds expense and complexity
Solution Approach 1:
The patent extracts the switching function from the backend display connection and moves it to the data transmission path. Instead of physically switching display connections, the system extracts and redirects scan-out data from the dGPU through the IGPU's raster generator, eliminating the need for backend switching mechanisms while maintaining display stability
Solution Approach 2:
The IGPU's raster generator serves as an intermediary that receives scan-out data from the dGPU and outputs it to the display. This intermediary approach allows the system to route graphics data without physical switching, maintaining continuous display operation while enabling GPU selection flexibility
2Reliability
If frame buffer selection is used to switch between IGPU and dGPU, then display disruption is avoided, but performance bottlenecks occur due to data copying across PCI-E bus and additional bandwidth consumption
Solution Approach 1:
The patent extracts only the necessary scan-out data from the dGPU's frame buffer and transmits it directly to the IGPU's raster generator, rather than copying entire frame buffers. This selective data extraction minimizes bandwidth consumption and avoids PCI-E bus bottlenecks while maintaining display stability
Solution Approach 2:
The system segments the graphics data transmission process by separating the rendering function (dGPU) from the display output function (IGPU raster generator). Only the scan-out data portion is transmitted between GPUs, rather than copying complete frame buffers, thereby reducing data transfer overhead and improving overall performance
3Productivity
If the dGPU handles all graphics tasks, then graphics performance is maximized, but power consumption increases and battery life decreases
Solution Approach 1:
The patent segments graphics task processing by routing intensive rendering tasks to the dGPU while handling display output through the low-power IGPU. This segmentation allows the system to leverage the dGPU's rendering capabilities without continuously consuming its high power, thereby reducing overall power consumption and extending battery life
Solution Approach 2:
The IGPU acts as an intermediary that receives rendered graphics data from the dGPU and handles the display output function. This intermediary role allows the system to use the low-power IGPU for display tasks while leveraging the dGPU only for rendering, optimizing the balance between graphics performance and power consumption
Data Source
AI summary
A method of displaying graphics data is described. The method involves accessing the graphics data in a memory subsystem associated with one graphics subsystem. The graphics data is transmitted to a second graphics subsystem, where it is displayed on a monitor coupled to the second graphics subsystem.