Fused Bounding Volume Hierarchy for Multi-LOD Cache Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing graphics rendering techniques face challenges in efficiently managing multiple levels of detail in bounding volume hierarchies, leading to increased data storage requirements and cache inefficiencies.
Innovation Solution
A fused bounding volume hierarchy is generated by merging base and non-base bounding volume hierarchies, where subtrees with fewer child nodes are identified and fused into the base hierarchy, reducing redundant data storage and improving cache efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple separate bounding volume hierarchies are used for different levels of detail, then each level can be optimized independently, but data storage requirements increase and cache efficiency decreases
Solution Approach 1:
The patent merges multiple separate bounding volume hierarchies (base Bvh and non-base Bvh) into a single fused bounding volume hierarchy. This combines the advantages of having optimized hierarchies for different levels of detail while eliminating the redundancy of storing separate data structures, thereby reducing overall data storage requirements while maintaining level-of-detail optimization capabilities
2Manufacturing precision
If multiple separate bounding volume hierarchies are used for different levels of detail, then each level can be optimized independently, but cache performance deteriorates due to increased data size
Solution Approach 1:
By merging multiple bounding volume hierarchies into a single fused structure, the patent reduces the total data size that needs to be loaded into cache memory. This improves cache performance and reduces cache misses while still maintaining the ability to efficiently render at different levels of detail through the fused hierarchy structure
3Quantity of substance
If a single bounding volume hierarchy is used, then data storage is efficient, but multiple levels of detail cannot be efficiently managed
Solution Approach 1:
The patent segments the single fused bounding volume hierarchy into different levels of detail by organizing nodes with level-of-detail identifiers. This allows the structure to efficiently manage multiple levels of detail while maintaining a compact single data structure, enabling selective traversal of appropriate detail levels based on rendering requirements
Data Source
AI summary
A fused bounding volume hierarchy, which is a combination of a base bounding volume hierarchy and one or more non-base bounding volume hierarchies, is generated. For each non-base bounding volume hierarchy, multiple subtrees in the non-base bounding volume hierarchy that include less than a threshold number of child nodes are identified. Each of these subtrees is then fused with the base bounding volume hierarchy at one of the nodes of the base bounding volume hierarchy, and an identifier of the level of detail for the non-base bounding volume hierarchy is included in the node. When displaying a scene or image, for a particular portion of the scene or image the level of detail to use is identified. The fused bounding volume hierarchy is traversed and the geometric objects in the nodes of the fused bounding volume hierarchy corresponding to the identified level of detail are displayed.