计算机应用 ›› 2010, Vol. 30 ›› Issue (2): 547-550.

• 典型应用 • 上一篇    下一篇

基于PC集群的三维图形并行渲染性能分析

彭帅1,李冬梅2,李朝晖2   

  1. 1. 中国传媒大学
    2.
  • 收稿日期:2009-08-30 修回日期:2009-09-13 发布日期:2010-02-10 出版日期:2010-02-01
  • 通讯作者: 彭帅

Parallel rendering efficiency analysis of 3D graphics on PC cluster

  • Received:2009-08-30 Revised:2009-09-13 Online:2010-02-10 Published:2010-02-01
  • Contact: shuai peng

摘要: 研究基于PC集群的三维图形并行渲染性能问题,从网络性能、算法复杂度、并行分配机制等几方面分析了影响并行渲染性能的关键因素。在千兆以太网PC集群上进行了基于通用MPI和OpenGL的三维图形并行渲染仿真测试,给出了数据及分析结果,给出了合理构建并行三维图形渲染系统的建议,通过平衡图形算法复杂度和网络性能以达到最佳并行性能。

关键词: 并行渲染, OpenGL, MPI, 并行效率

Abstract: In this paper, the efficiency of parallel 3D graphics rendering on general-purpose PC cluster was investigated. Three key factors affecting the efficiency of parallel rendering were analyzed, including network performance, algorithm complexity and parallel rendering mechanism. Parallel rendering simulations based on MPI and OpenGL were carried out on gigabit Ethernet PC cluster. The experimental results were analyzed and the advices for designing efficient parallel rendering system on PC cluster were given.

Key words: parallel rendering, OpenGL, MPI, parallel efficiency