Massively Parallel Hierarchical Scene Sorting with Applications in Rendering

We present a novel method for massively parallel hierarchical scene processing on the GPU, which is based on sequential decomposition of the given hierarchical algorithm into small functional blocks. The computation is organized using a specialized work pool in which different blocks of processing units solve different functional blocks. We present an application of the proposed approach to two methods used in ray tracing: construction of the bounding volume hierarchies and the recently introduced divideand-conquer ray tracing on the GPU. The results indicate that using our approach we achieve high utilization of the GPU even for complex hierarchical problems which pose a challenge for massive parallelization.