海量数据GPU体绘制优化研究  

Optimizing Large Scale Data GPU Volume Rendering

作　　者：苏超轼[1] SU Chao-shi　

机构地区：[1]桂林电子科技大学,计算机与控制学院,广西,桂林,541004

出　　处：《电脑知识与技术（过刊）》2007年第20期437-,463,共2页Computer Knowledge and Technology

摘　　要：高性能GPU使得体绘制在廉价的硬件上获得良好的性能,但海量数据体绘制的效率依旧低下.本文探讨了GPU体绘制中图形硬件的瓶颈,并提出新颖的算法解决这些问题:采用数据分块和八叉树划分体数据实现空单元跳过优化.该算法解决了海量数据超过可用纹理空间的难题,同时允许实时改变体绘制传递函数.Volume Rendering methods employing the GPU capabilities offer high performance on available hardware, but large scale data volume rendering also less efficient. In this article, we discuss the various bottlenecks in the graphics hardware when performing GPU-based Volume Rendering. Further An novel algorithm is proposed to overcome these bottlenecks existing in rendering large scale volume data: we apply Octree and bricks to divide the volume data for empty space skipping optimization. The described method takes into account that the rendered volume may exceed the available texture memory and further allows fast run-time changes of the transfer function.

关 键 词：体绘制 图形处理器 八叉树 数据分块 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]