Answers are grounded in this paper and cite their sources.
We analyze and compare the performances of two acceleration data structures for electromagnetic ray-tracing purposes on graphical processing units (GPUs) using the CUDA programming language, namely the K-Dimensional (KD)-tree and the Split Bounding Volume Hierarchy (SBVH). Our implementations have been based on the approach made available by Nvidia, which takes into account the programming optimizations made possible by the latest version of CUDA and the most recent Nvidia GPU architectures. We have tested the two approaches on standard computer graphics scenes (conference and bunny) and on a scene of electromagnetic interest ("ship"). In all the cases considered, the SBVH has shown to perform better in terms of both speed and memorysaving properties.
No takes yet. Share an insight, caveat, or question.
Breglia et al. (2015) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: