计算机应用 ›› 2011, Vol. 31 ›› Issue (02): 530-532.

• 先进计算与信号处理 • 上一篇    下一篇

网格作业完工时间与作业分割粒度的关系

李荣胜1,赵文峰2,徐惠民2   

  1. 1. 北京邮电大学
    2.
  • 收稿日期:2010-07-19 修回日期:2010-09-17 发布日期:2011-02-01 出版日期:2011-02-01
  • 通讯作者: 李荣胜
  • 基金资助:
    国家973计划项目;贵州省重大科技专项计划

Relationship between makespan of grid job and granularity of job partitioning

  • Received:2010-07-19 Revised:2010-09-17 Online:2011-02-01 Published:2011-02-01

摘要: 对可分割的计算密集型大型作业在并行且不间断运行情况下的完工时间与作业分割粒度之间的关系进行研究。首先分析了子作业之间无通信和有通信两种情况下可分割计算密集型大型作业的完工时间和分割粒度的关系,然后对可分割计算密集型大型作业在专用网格资源上的完工时间与分割粒度的关系进行仿真。仿真结果显示,大型作业的完工时间随着分割粒度的增大先减小后增大;当单个子作业的计算时间和通信时间之比增大时,作业的分割粒度可以更细,作业完工时间的最小值减小。因此完工时间最优的作业分割粒度不能过粗或过细。

关键词: 完工时间, 分割粒度, 计算密集型, 作业调度, 网格计算

Abstract: There have been many divisible computeintensive grand-challenge jobs running on volunteer grid. The relationship between makespans of such grid jobs and granularities of jobs partitioning was studied. Firstly, the relationship between makespan and partitioning granularity of computeintensive jobs with and without communications between subjobs were analyzed theoretically. Then, the relationship between makespan and partitioning granularity of a job with and without communications between subjobs running on dedicated grid resources in parallel mode were simulated. The simulation results show that grandchallenge jobs makespan decreases at first and then increases when granularity increases. Granularity can be more grain, and the best makespan will decrease when the ratio of computation time to the communication time of a subjob increases. To optimize jobs makespan, the job's partitioning granularity should not be too coarse or too fine.

Key words: makespan, partitioning granularity, compute-intensive, job scheduling, grid computing