一种基于GPU加速的三维分形实时渲染方法

A real time drawing method for three-dimensional fractal image based on GPU acceleration

绘制高质量三维分形图所需的计算量往往很大,难以达到实时的要求。研究了一种绘制三维分形的通用方法,即基于光线跟踪的距离估计算法。该算法非常适合在GPU上实现,利用GPU强大的并行处理能力,使三维分形的绘制速度显著提高。大量实验结果表明,在屏幕分辨率为1280x720时绝大部分三维分形图都已达到了实时绘制的要求。GPU与CPU的性能对比表明,即使在CPU最快的情况下耗时也是GPU的158倍,由此可见应用GPU的巨大优势。还验证了距离估计算法的通用性,对于不同类型的分形,如逃逸时间算法的分形、迭代函数系统的分形以及分形地形都可以适用。

The volumes of calculation in drawing high-quality three-dimensional fractal images are often huge. Therefore, it is difficult to meet the requirement for real time drawing. A universal method for drawing 3-D fractal im- age is studied in this article. That is the Distance Estimation （DE） Algorithm based on ray tracing. The DE algorithm is very suitable to be realized on GPU. And it takes the advantage of the strong parallel processing power of GPU, which significantly increases the speed of 3-D fractal drawing. Many experimental results show that the vast majority of 3-D fractal images drawn under the 1,280 × 720 screen resolution have met the requirement for real time drawing. The performance comparison between GPU and CPU shows that even in the case of the fastest CPU, the consumed time of CPU is 158 times more than that of GPU. Thus it can be seen the great advantage of using GPU. The generality of the DE algorithm is also verified in this article, that is, the DE algorithm is applicable to different kinds of fractal images such as Escape Time Algorithm, Iterated Function System and Fractal Terrain.