Studying the Effect of Schedule Preemption on Dynamic Task Graph Scheduling
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908808082620416 |
|---|---|
| author | Khodabandehlou, Mohammadali Coleman, Jared Suri, Niranjan Krishnamachari, Bhaskar |
| author_facet | Khodabandehlou, Mohammadali Coleman, Jared Suri, Niranjan Krishnamachari, Bhaskar |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_03081 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Studying the Effect of Schedule Preemption on Dynamic Task Graph Scheduling Khodabandehlou, Mohammadali Coleman, Jared Suri, Niranjan Krishnamachari, Bhaskar Distributed, Parallel, and Cluster Computing C.2.4; C.2.1 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. |
| title | Studying the Effect of Schedule Preemption on Dynamic Task Graph Scheduling |
| topic | Distributed, Parallel, and Cluster Computing C.2.4; C.2.1 |
| url | https://arxiv.org/abs/2602.03081 |