A Comparison of GPU Box-Plane Intersection Algorithms for Direct Volume Rendering

Stefan Zellmann and Ulrich Lang

Keywords

Volume Rendering, Rendering Algorithms and Systems, BoxPlane Intersection Algorithms, Geometry Programs

Abstract

In order to avoid load imbalances on the GPU during direct volume rendering, a common scheme was to move the generation of proxy geometry, that quite often consists of polygons retrieved through box-plane intersections, from the CPU to the vertex stage of the GPU. Nowadays, with the unified shader architectures implemented by modern graphics hardware, usually non of the programmable stages will starve anymore. Nevertheless, redistributing calculations for proxy geometry generation from the vertex stage to the newly introduced geometry stage of the graphics hardware results in some serial computations that can be performed more finely grained, thus hinting the graphics driver to schedule more tasks in parallel. We propose different implementations of box-plane intersection algorithms on the GPU that use the vertexas well as the geometry stage and compare their performance on modern graphics hardware.

Important Links:



Go Back