Tiled Graphics Processor Adaptive Tiling Unit Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tile-based graphics processing systems have limitations in efficiency and configurability, particularly due to the use of a single tiling unit, which can lead to suboptimal performance in handling complex rendering tasks and varying workloads.
Innovation Solution
Implementing a graphics processor with multiple tiling units that can operate independently, allowing for adaptive allocation of tiling units based on the complexity of the rendering task, and employing a two-stage sorting process involving initial 'hypertiling' and subsequent tiling to efficiently process geometry into primitive listing regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single tiling unit is used in tile-based graphics processing, then the device complexity is reduced, but the productivity and processing efficiency deteriorate when handling complex rendering tasks
Solution Approach 1:
The render output is divided into initial regions (hypertiles) that encompass multiple primitive listing regions, allowing a single tiling unit to process multiple regions sequentially without requiring multiple simultaneous tiling units
Solution Approach 2:
The system dynamically allocates tiling units to different initial regions based on workload characteristics, enabling adaptive processing efficiency without permanently increasing device complexity
2Productivity
If multiple tiling units are deployed to handle complex rendering tasks, then the productivity improves, but the device complexity and resource allocation difficulty increase
Solution Approach 1:
The system provides dynamic allocation of tiling units to initial regions, allowing the configuration to adapt automatically to different rendering workloads without requiring manual intervention or complex control mechanisms
Solution Approach 2:
Each tiling unit is designed to be multi-functional, capable of processing different types of geometry (triangles, lines, points) and handling various rendering scenarios, reducing the need for specialized units
3Manufacturing precision
If geometry is sorted into fine-grained primitive listing regions, then the manufacturing precision of rendering is improved, but the loss of time increases due to repeated processing
Solution Approach 1:
The system performs preliminary sorting of geometry into initial regions (hypertiles) before detailed primitive listing, preventing duplication of sorting work and reducing overall processing time while maintaining rendering precision
Solution Approach 2:
The hierarchy of regions (initial regions containing multiple primitive listing regions) creates a nested structure where sorting at the higher level (initial regions) eliminates the need for redundant sorting at lower levels, saving processing time
Data Source
AI summary
A tiled-based graphics processor that comprises a plurality of tiling units is disclosed. The graphics processor includes an assigning circuit that assigns tiling units to sort geometry for initial regions of a render output that encompass plural primitive listing regions, and causes assigned tiling units to sort geometry for an initial region into primitive listing regions that the initial region encompasses.