arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.03081cs.DC
收藏

研究调度抢占对动态任务图调度的影响

Studying the Effect of Schedule Preemption on Dynamic Task Graph Scheduling

University of Southern California(南加州大学);Loyola Marymount University(洛约拉马利蒙特大学);US Army Research Laboratory(美国陆军研究实验室)

Mohammadali Khodabandehlou, Jared Coleman, Niranjan Suri, Bhaskar Krishnamachari

AI总结:

本文提出Last-K抢占模型,通过适度抢占提升动态任务图调度的完成时间和利用率,同时保持公平性和低开销。

摘要:

动态调度任务图通常不重新审视先前的任务分配,主要关注最小化完成时间。我们研究受控的调度抢占,引入了Last-K抢占模型,该模型选择性地重新调度最近的任务图,同时保留早期的分配。使用合成、RIoTBench、WFCommons和对抗性工作负载,我们在完成时间、公平性、利用率和运行时间方面比较抢占性、非抢占性和部分抢占性策略。结果表明,适度的抢占可以实现与完全抢占相似的完成时间和利用率提升,同时保持公平性和低开销。

英文摘要:

Dynamic scheduling of task graphs is often addressed without revisiting prior task allocations, with a primary focus on minimizing makespan. We study controlled schedule preemption, introducing the Last-K Preemption model, which selectively reschedules recent task graphs while preserving earlier allocations. Using synthetic, RIoTBench, WFCommons, and adversarial workloads, we compare preemptive, non-preemptive, and partial-preemptive strategies across makespan, fairness, utilization, and runtime. Results show moderate preemption can match most makespan and utilization gains of full preemption while maintaining fairness and low overhead.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑