Parallel BVH Reinsertion for Non-Binary Ray Tracing Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for optimizing bounding volume hierarchies (BVHs) in ray tracing, such as Meister's approach, are limited by the requirement to maintain a binary tree structure, which restricts the potential for reducing the surface area heuristic (SAH) and may not yield the most efficient computational cost reductions.

Innovation Solution

A method that allows for non-binary reinsertions in the BVH, where nodes can have more than two children, and employs a deterministic allocation scheme using process IDs to manage node slots, along with conflict retries and restricted search scopes to enhance computational efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If binary tree structure is maintained in BVH optimization, then structural simplicity and ease of implementation are improved, but SAH reduction potential and computational efficiency deteriorate

Engineering Contradiction:
Improveease of implementationVSAvoidcomputational efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent changes the structural parameter of the BVH from strictly binary to non-binary, allowing nodes to have varying numbers of children. This parameter change enables greater flexibility in optimizing the hierarchy structure to reduce SAH, while still maintaining implementation feasibility through systematic node reinsertion algorithms and deterministic slot allocation schemes.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If non-binary reinsertions are allowed in BVH, then SAH savings and computational cost reduction are improved, but structural complexity and implementation difficulty increase

Engineering Contradiction:
ImproveSAH savingsVSAvoidstructural complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements self-service mechanisms where the BVH optimization process automatically manages its own complexity through deterministic slot allocation based on process IDs. Each reinsertion operation self-manages slot assignment and structure updates without requiring external coordination, enabling non-binary structures to be handled systematically despite increased complexity.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent introduces a new structural parameter allowing variable child counts per node in the BVH hierarchy. This parameter change from fixed binary to flexible non-binary structure enables better SAH optimization by adapting the tree topology to the specific spatial distribution of primitives, achieving up to 10% better performance in certain scenarios.

Inventive Principle:
Principle #35Parameter changes

3Speed

If parallel reinsertion processes are used, then optimization speed is improved, but determinism and result consistency deteriorate

Engineering Contradiction:
Improveoptimization speedVSAvoidresult consistency
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent introduces deterministic slot allocation as an intermediary mechanism that mediates between parallel reinsertion processes. By using process IDs to deterministically assign and manage node slots, the system coordinates parallel operations without requiring complex inter-process communication or synchronization, maintaining both speed and determinism.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

Each parallel reinsertion process serves itself by deterministically selecting and managing its own node slots based on its process ID. This self-service approach eliminates the need for inter-process coordination while ensuring that all processes operate on consistent data, maintaining determinism across parallel executions.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4254344B1Parallel optimisation of bounding volume hierarchy
Publication Date: 2025.07.09 IMAGINATION TECH LTD
  • EP4254344B1 patent drawingFigure 1~2
  • EP4254344B1 patent drawingFigure 3a~4b
  • EP4254344B1 patent drawingFigure 5

AI summary

A method comprising, in each of one or more iterations: a) for each of a plurality of input nodes in a current bounding volume hierarchy (BVH) searching for one or more candidate reinsertions, which would move the input node from a respective old parent to a respective new parent, and which compared to the current BVH would reduce an expected computational cost of searching the tree for a ray intersection; and b) updating the current BVH with one or more selected reinsertions from among the beneficial candidate reinsertions. In the search of at least one iteration, the one or more candidate reinsertions include at least one candidate reinsertion that would leave the old parent with more than one child, or give the new parent more than two children.