计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (16): 176-178.

• 图形、图像、模式识别 • 上一篇    下一篇

利用Master-Slave-Collector模式的大规模数据集的并行体绘制

汤 敏   

  1. 南通大学 电子信息学院,江苏 南通 226007
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-06-01 发布日期:2011-06-01

Parallel volume rendering of large-scale datasets based on Master-Slave-Collector mode

TANG Min   

  1. College of Electronics and Information,Nantong University,Nantong,Jiangsu 226007,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-06-01 Published:2011-06-01

摘要: 以内部网络和普通配置计算机为实验平台,研究大规模数据集的并行体绘制的实现方法,以提高绘制速度和算法效率。分别介绍并行可视化、Master-Slave-Collector模式、负载平衡、任务池和结果池等关键技术。在传统的Master-Slave模式基础上的改进模式Master-Slave-Collector,具有减少计算时间、实现负载平衡、提高绘制效率等优点。实验结果表明,该方法较好地解决了运算速度和内存空间这两大难题,效果良好,实时性强,在临床诊断和科学研究中发挥重要作用。

关键词: 并行体绘制, 机群系统, 科学计算可视化

Abstract: The parallel strategy that can effectively improve the rendering speed and the algorithm efficiency is proposed on the basis of Intranet and common-configuration computers. Several kernel techniques are introduced respectively, including the basic concepts of parallel volume rendering and visualization, Master-Slave-Collector mode, load balance, pool-of task and pool-of result. The Master-Slave-Collector computation mode presented can decrease computing time, attain load balance, and improve the rending speed effectively without reducing image quality. Experimental results demonstrate that this method provides promising and real-time results to play great role in clinic and research field, which balances the computational speed and memory requirements.

Key words: parallel volume rendering, cluster, visualization in scientific computation