GPU Tile Rendering Pipelines for Ordered Primitive Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing graphics processing units (GPUs) face challenges in reducing latency, power consumption, and size while implementing tile-based rendering techniques, often requiring trade-offs between these factors.

Innovation Solution

A graphics processing unit (GPU) is configured to process primitives using a rendering space divided into tiles, featuring a geometry processing phase with parallel geometry pipelines and a tiling module that includes tiling front-end and back-end modules, allowing for efficient determination and ordering of tile-primitive associations, and a dynamic allocation scheme for tile pipelines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If tile-based rendering techniques are implemented in GPU, then rendering efficiency is improved, but device complexity increases

Engineering Contradiction:
Improverendering efficiencyVSAvoiddevice complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The rendering space is divided into multiple tiles, and the geometry processing logic is segmented into multiple geometry pipelines that can process different tiles in parallel. Each pipeline independently handles tiling operations for its assigned tiles, enabling concurrent processing and improving rendering efficiency while distributing complexity across multiple simpler units

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically allocates tiles to geometry pipelines based on workload characteristics. The tiling module can adaptively determine which geometry pipeline processes which tiles, allowing the system to optimize performance for different rendering scenarios and primitive distributions without requiring a fixed complex architecture

Inventive Principle:
Principle #15Dynamics

2Productivity

If multiple geometry pipelines are used for parallel processing, then productivity is improved, but loss of time due to coordination overhead increases

Engineering Contradiction:
ImproveproductivityVSAvoidcoordination overhead
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary tiling operations within each geometry pipeline before final control stream generation. By determining tile-primitive associations early in the geometry processing phase and caching intermediate results, the system reduces the coordination overhead required during fragment processing, allowing parallel pipelines to work independently with minimal synchronization

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The tiling module acts as an intermediary that receives tile-primitive indications from multiple geometry pipelines and consolidates them into unified control streams. This mediator coordinates the output from parallel pipelines efficiently, managing the aggregation and ordering of primitive data without requiring complex direct communication between pipelines

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If tiling module determines tile-primitive associations for all tiles, then measurement precision is improved, but use of energy increases

Engineering Contradiction:
Improvetiling accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

Each geometry pipeline performs tiling operations with high precision for its assigned tiles, ensuring accurate tile-primitive associations locally. The system maintains measurement precision by dedicating sufficient computational resources to each pipeline's specific tile set, while the overall energy consumption is managed by distributing work across multiple pipelines that can operate independently and efficiently

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4645233A1Graphics processing unit for processing primitives using a rendering space which is sub-divided into a plurality of tiles
Publication Date: 2025.11.05 IMAGINATION TECH LTD
  • EP4645233A1 patent drawingFigure 1
  • EP4645233A1 patent drawingFigure 2
  • EP4645233A1 patent drawingFigure 3

AI summary

A graphics processing unit (GPU) and method are described herein for processing a sequence of primitives using a rendering space which is sub-divided into a plurality of tiles. The GPU comprises geometry processing logic which comprises a plurality of geometry pipelines and a tiling back-end module. The geometry pipelines are configured to receive batches of primitives of the sequence of primitives. Each of the geometry pipelines comprises: one or more geometry processing modules configured to perform one or more geometry processing functions on the primitives of a batch of primitives received at the geometry pipeline; and a tiling front-end module configured to determine, for each tile of a set of one or more tiles, one or more tile-primitive indications indicating which of the primitives of the batch of primitives received at the geometry pipeline are present within that tile. The tiling back-end module is configured to: receive the tile-primitive indications determined by the plurality of geometry pipelines; and for each of the tiles for which a tile-primitive indication is received, include indications of the primitives that are present within that tile in a control stream for that tile in an order in accordance with an order of the primitives within the sequence of primitives.