Resumo

LE’CLERC ARRASTIA, Jean; PIAD MORFFIS, Alejandro  e  LEONARD MENDEZ, Ludwig. A-PIT: Spatial Subdivision Structure for Raytracing Acceleration on GPU. Rev cuba cienc informat [online]. 2016, vol.10, n.3, pp. 71-87. ISSN 2227-1899.

ABSTRACT The digital image synthesis has a broad range of applications, which have stimulated research in hardware and algorithms to this end. The raytracing algorithm has been widely used for this purpose, especially to simulate certain physical phenomena that arise in the interaction between light and matter. Its simplicity and elegance enable the simulation of optical effects such as reflection, refraction and shadows. Detecting the first object that intersects a ray, which is the basis of this algorithm, is a computationally complex operation that can easily consume 95% of rendering time. Hence, several data structures have been designed to reduce the complexity of this operation. This paper proposes a new data structure, the A-PIT, designed to accelerate the ray intersection operation in a GPU implementation of raytracing. We show that using this data structure in an accumulative version of raytracing enables the interactive rendering of scenes of medium complexity.

Palavras-chave : raytracing; gpu; a-buffer; interval tree; subdivision.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )