Primitive ID Map Sampling for Accurate Small-Triangle Lookup

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing three-dimensional graphics processing systems face challenges in accurately mapping points in a texture space to primitives in a three-dimensional object space due to imprecise primitive ID maps, leading to increased memory requirements and potential lookup failures, especially for small triangles.

Innovation Solution

The implementation of a primitive ID map that exploits spatial redundancy by assigning redundant primitive ID samples to unrepresented triangles, combined with a neighborhood search mechanism to ensure precise lookup, thereby improving the mapping process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a high-resolution primitive ID map is used to ensure accurate mapping of texture samples to primitives, then lookup precision is improved, but memory requirements increase

Engineering Contradiction:
Improvelookup precisionVSAvoidmemory requirements
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the primitive ID map into multiple tiles or regions, allowing the system to load only the necessary portions into memory during rendering. This division enables accurate lookup precision for visible regions while reducing overall memory requirements by not loading the entire high-resolution map at once.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a hierarchical structure where a low-resolution primitive ID map provides overall coverage, and high-resolution details are nested within specific regions as needed. This nested approach allows the system to maintain lookup precision for important areas while using minimal memory for the base structure.

Inventive Principle:
Principle #7Nested doll (Nesting)

2Quantity of substance

If a low-resolution primitive ID map is used to reduce memory requirements, then memory usage decreases, but lookup accuracy deteriorates leading to mapping failures

Engineering Contradiction:
Improvememory usageVSAvoidmapping accuracy
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent applies local quality by assigning different resolution levels to different regions of the primitive ID map based on their importance and visibility. Regions that require high mapping accuracy (such as areas with small triangles or complex geometry) are rendered at high resolution, while less critical areas use lower resolution, thereby maintaining reliability where needed while reducing overall memory usage.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent introduces a third dimension of quality by storing multiple resolution levels or quality tiers for different regions. This dimensional approach allows the system to select the appropriate resolution level based on the specific rendering context, ensuring mapping accuracy is maintained for critical regions while optimizing memory usage across the entire scene.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Reliability

If redundant primitive ID samples are assigned to unrepresented triangles, then coverage of all primitives is improved, but data structure complexity increases

Engineering Contradiction:
Improveprimitive coverageVSAvoiddata structure complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary action by pre-identifying triangles that would otherwise be unrepresented in the primitive ID map and proactively assigning redundant primitive ID samples to them during the map generation phase. This advance preparation ensures complete primitive coverage during rendering without requiring complex runtime detection or correction mechanisms.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses copying by creating redundant copies of primitive ID samples and assigning them to triangles that would otherwise lack representation. These copied samples are placed in strategic locations within the primitive ID map structure, ensuring that all primitives can be reliably mapped without requiring fundamental changes to the data structure complexity.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS12511815B2System and method for primitive ID map sampling
Publication Date: 2025.12.30 ADVANCED MICRO DEVICES INC
  • US12511815B2 patent drawing
  • US12511815B2 patent drawing
  • US12511815B2 patent drawing

AI summary

A technique for sampling a primitive ID map. The technique includes identifying a sample point having a location in a texture space; obtaining a primitive ID sample from the primitive ID map based on the location of the sample point in the texture space; identifying a primitive based on the primitive ID; testing the location in the texture space for inclusion within the identified primitive; and selecting either the primitive ID or a different primitive ID based on the testing.