Masked Depth Culling for Graphics Hardware

摘要

Hierarchical depth culling is an important optimization, which isrnpresent in all modern high performance graphics processors. Wernpresent a novel culling algorithm based on a layered depth representation,rnwith a per-sample mask indicating which layer eachrnsample belongs to. Our algorithm is feed forward in nature in contrastrnto previous work, which rely on a delayed feedback loop. Itrnis simple to implement and has fewer constraints than competingrnalgorithms, which makes it easier to load-balance a hardware architecture.rnCompared to previous work our algorithm performs veryrnwell, and it will often reach over 90% of the efficiency of an optimalrnculling oracle. Furthermore, we can reduce bandwidth by up torn16% by compressing the hierarchical depth buffer.