作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (22): 258-260. doi: 10.3969/j.issn.1000-3428.2008.22.091

• 开发研究与设计技术 • 上一篇    下一篇

图形处理器的流执行模型

李海燕,张春元,李 礼,任 巨   

  1. (国防科技大学计算机学院,长沙 410073)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-11-20 发布日期:2008-11-20

Stream Execution Model of Graphics Processing Unit

LI Hai-yan, ZHANG Chun-yuan, LI Li, REN Ju   

  1. (School of Computer, National University of Defense Technology, Changsha 410073)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-11-20 Published:2008-11-20

摘要: 图形处理器极高的流计算能力使其成为实现实时流应用的有效方案。该文抽象出图形处理器的流执行模型,描述图形处理器流处理机制的执行过程,在图形处理器上实现了二维离散余弦变换。实验结果表明,图形处理器对标清格式的视频压缩编码效率可达70 fps。

关键词: 图形处理器, 流处理, 流执行模型

Abstract: Graphics Processing Unit(GPU), which has powerful stream computing capability, makes itself as an efficient scheme for real-time stream applications. This paper abstracts a stream execution model of GPU, describes its stream execution process, and implements the discrete cosine transform on GPU. Experimental results show that the coding efficiency of standard definition video compression on GPU can reach 70 fps.

Key words: Graphics Processing Unit(GPU), stream processing, stream execution model

中图分类号: