Variable Rasterization Order for Motion Blur Cache Thrashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current graphics processing techniques for rendering motion blur and depth of field in real-time graphics face inefficiencies due to unpredictable performance and poor scaling with increased blur, especially with smaller primitives, leading to cache thrashing and strobing artifacts.
Innovation Solution
Reversing the traversal order of triangles in the rasterization pipeline, either for each successive triangle or group of triangles, to improve cache coherence and reduce cache thrashing, while disabling reversal for small motion or specific geometric configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If stochastic sampling is used for motion blur and depth of field, then rendering quality is improved, but performance becomes unpredictable and scales poorly with increased blur
Solution Approach 1:
The patent inverts the conventional approach by using deterministic rasterization order (reverse order) instead of stochastic sampling. This reversal maintains rendering quality while achieving predictable performance by systematically traversing triangles in reverse order rather than using random sampling patterns.
Solution Approach 2:
The patent changes the traversal order parameter from standard to reverse order, and dynamically adjusts this parameter based on motion magnitude thresholds. This parameter change enables the system to switch between different rasterization strategies to optimize both quality and performance predictability.
2Manufacturing precision
If smaller primitives are used, then geometric detail is improved, but rasterization efficiency deteriorates due to increased traversal overhead
Solution Approach 1:
The patent applies reverse triangle traversal order specifically optimized for small primitives. By traversing triangles in reverse order, the system reduces redundant screen-space traversal for small geometric details, thereby maintaining high geometric detail while improving rasterization efficiency.
3Productivity
If accumulation buffering or interleaved sampling is used, then constant cost is achieved, but strobing artifacts are introduced
Solution Approach 1:
The patent uses reverse triangle traversal order as an alternative to accumulation buffering and interleaved sampling. This inverted approach achieves constant cost performance while avoiding strobing artifacts by systematically processing triangles in reverse order rather than using accumulation techniques.
4Device complexity
If standard triangle traversal order is used, then rendering simplicity is maintained, but cache coherence deteriorates leading to cache thrashing
Solution Approach 1:
The patent inverts the triangle traversal order to improve cache coherence. By rendering triangles in reverse order, the system achieves better spatial and temporal locality in memory access patterns, reducing cache thrashing while maintaining relatively simple implementation.
5Manufacturing precision
If motion blur and depth of field are rendered with small primitives, then scene detail is improved, but the amount of work performed increases excessively
Solution Approach 1:
The patent uses reverse triangle traversal to reduce the screen-space traversal area for small primitives. By traversing in reverse order, the system minimizes redundant work performed on blurred primitives, thereby reducing computational energy consumption while maintaining high scene detail.
Data Source
AI summary
Cache thrashing or over-accessing of a cache can be reduced by reversing the order of traversal of a triangle on different granularities. In the case where triangles are not grouped, the traverse order may be reversed on each triangle. In cases where triangles are grouped, the traversal order may be reversed with each group change. However, when motion is excessive, for example beyond a threshold, then the traversal order reversal may be disabled.