启发式任务调度中的处理器选择策略

任务调度是并行分布计算中最为基本、最为关键,也最具有挑战性的问题之一,是影响并行分布计算执行效率的一个关键因素.现有的基于任务静态优先级的启发式任务调度方法都是以“当前任务具有最早起始执行时刻”为目标来选择执行处理器.该文在详细分析讨论该种调度方法的基础上,指出了以该目标选择处理器存在的问题及缺点,并提出了以“当前任务的直接后继具有最早起始执行时刻”为目标选择处理器的方法,并给出了相应的约束条件.

Processor Selection Policy in Heuristic Task Scheduling

As one of the most fundmental, critical and challengable problems in PDC (parallel distributed computing), task scheduling has great influence on the execution efficiency of PDC. The existing heuristic task schedulings based on static task priorities always use it as processor selection policy to make the current task have earliest start execution time. On the basis of analysis of the mechanism in priority-based heuristic task scheduling, the authors illustrate the drawbacks of the processor selection policy mentioned above, and propose a new processor selection policy, i.e, to make the successor of the current task have the earliest start time, and give the corresponding restraint.