GPU ray tracing based on reduced bounding volume hierarchies

Sai-Keung Wong*, Yu Chun Cheng, Shing Yeu Lii

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

We present a bounding volume hierarchy construction based on a culling technique for performing ray tracing on manycore GPUs. The objects are divided into view-sets of triangles with respect to a set of view targets (including the camera and the light sources). A bounding volume hierarchy, namely reduced bounding volume hierarchy (RBVH), is then constructed for each view-set. Each RBVH contains a portion of the triangles of the objects. The ray traversal is improved for primary and shadow rays, leading to a net improvement of ray tracing. We also handle multiple lights that are close to each other. Our method has been implemented on manycore GPUs and evaluated for dynamics scenes. Experimental results show that our method improves ray tracing for primary and shadow rays.

Original languageEnglish
Title of host publicationCGiV 2012 - 2012 9th International Conference on Computer Graphics, Imaging and Visualization
Pages1-6
Number of pages6
DOIs
StatePublished - 6 Nov 2012
Event2012 9th International Conference on Computer Graphics, Imaging and Visualization, CGiV 2012 - Hsinchu, Taiwan
Duration: 25 Jul 201227 Jul 2012

Publication series

NameCGiV 2012 - 2012 9th International Conference on Computer Graphics, Imaging and Visualization

Conference

Conference2012 9th International Conference on Computer Graphics, Imaging and Visualization, CGiV 2012
Country/TerritoryTaiwan
CityHsinchu
Period25/07/1227/07/12

Keywords

  • bounding volume hierarchy
  • culling technique
  • manycore GPUs
  • ray tracing

Fingerprint

Dive into the research topics of 'GPU ray tracing based on reduced bounding volume hierarchies'. Together they form a unique fingerprint.

Cite this