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.

翻译：动态任务图调度通常在不重新审视先前任务分配的情况下进行，主要侧重于最小化完工时间。我们研究受控调度抢占，引入Last-K抢占模型，该模型选择性地重新调度最近的任务图，同时保留较早的分配。使用合成、RIoTBench、WFCommons和对抗性工作负载，我们在完工时间、公平性、利用率和运行时间方面比较抢占式、非抢占式和部分抢占式策略。结果表明，适度抢占可以在保持公平性和低开销的同时，匹配完全抢占的大部分完工时间和利用率收益。