计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (11): 150-153.

• 图形图像处理 • 上一篇    下一篇

基于ParaViewWeb架构的GPU高性能运算实现

褚晶辉,王亚琦,吕  卫   

  1. 天津大学 电子信息工程学院,天津 300072
  • 出版日期:2015-06-01 发布日期:2015-06-12

GPU high-performance computing implementation based on ParaViewWeb framework

CHU Jinghui, WANG Yaqi, LV Wei   

  1. School of Electronic Information Engineering, Tianjin University, Tianjin 300072, China
  • Online:2015-06-01 Published:2015-06-12

摘要: 针对智能手机、平板电脑等移动终端计算能力有限的问题,提出一种基于网络的通用计算服务平台。该平台以服务的方式为移动终端提供远程计算支持,增强其进行复杂运算的能力。利用ParaViewWeb架构的可扩展性,将高性能GPU通用计算与ParaViewWeb架构结合,在保留ParaViewWeb原有的三维数据分析与可视化功能的基础上,扩展其远程调用服务器GPU运算资源的能力。选取医学图像滤波算法为实验对象,验证了该平台的有效性,结果表明对比本地应用,基于网络的应用取得了显著加速。

关键词: ParaViewWeb, 统一计算设备架构, 高性能计算, 移动设备

Abstract: This paper presents a network-based general computing platform which provides computation as a service to enable the mobile terminals such as smart phones and tablet PCs to do complex computational work. Based on the expandability of ParaViewWeb’s framework, capabilities of high-performance general computing with GPUs are added while the original functions of 3-D data analysis and visualization are kept. Now the GPU computing resources of the server can be called remotely on this platform. The platform is tested with a medical image filtering application and the experimental result shows the significant acceleration of the web-based application compared to the locally running application.

Key words: ParaViewWeb, compute unified device architecture, high-performance computing, mobile devices