计算机应用 ›› 2010, Vol. 30 ›› Issue (2): 543-546.

• 典型应用 • 上一篇    下一篇

使用GPU技术的数据流分位数并行计算方法

周勇1,王皓2,程春田3   

  1. 1. 大连理工大学
    2. 大连理工大学软件学院
    3.
  • 收稿日期:2009-08-02 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: 王皓
  • 基金资助:
    电力市场环境下省级电网水火电协调竞争优化及风险分析方法

Parallel computing method of data stream quantiles with GPU

  • Received:2009-08-02 Revised:2009-09-13 Online:2010-02-10 Published:2010-02-01
  • Contact: Hao Wang

摘要: 数据流实时、连续、快速到达的特点决定了数据流的实时处理能力。在处理低维数据流时经常使用分位数信息来描述数据流的统计信息,利用图形处理器(GPU)的强大计算能力和高内存带宽的特性计算数据流分位数信息,提出了基于统一计算设备架构(CUDA)的数据流处理模型和基于该模型的数据流分位数并行计算方法。实验证明,该方法在提供不低于纯CPU分位数算法相同精度的条件下,使数据流分位数的实时计算带宽得到了显著的提高。

关键词: 统一计算设备架构, 通用图形处理器, 数据流, 分位数, 并行计算

Abstract: The real-time, continuous and rapid arrival properties of data streams decide the real-time processing capability of data stream. Quantiles are commonly used for describing data stream with low dimension distribution. The research focused on mining powerful computing capacity and high memory bandwidth of Graphics Processing Unit (GPU) to compute data stream quantiles, and presented a GPU cooperated parallel processing model of data stream based on Computing Unified Device Architecture (CUDA) as well as parallel computing method of data stream quantiles which increased data stream processing bandwidth remarkably with precision no less than pure CPU algorithm.

Key words: Computing Unified Device Architecture (CUDA), general-purpose computing on graphics processing unit, data stream, quantile, parallel computing