EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

FlePo: GPU Multitask Scheduling Optimization Framework for Dynamic Scenes

Huanghuang Liang, Xin Yang, Rui Ge, Yaqi Xia, Chuang Hu, Xiaobo Zhou, Dazhao Cheng

ACM Transactions on Autonomous and Adaptive Systems · 30.09.2026 · 21 (3) · 1-35

DOI: 10.1145/3807775 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Autonomous and Adaptive Systems, online erschienen am 10.06.2026. Verfasst von Huanghuang Liang und 6 weiteren Autor:innen. Die Arbeit nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, National Key Research and Development Program of China, Science and Technology Development Fund, Macao S.A.R (FDCT) Projects.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep Neural Networks (DNNs) are widely used in intelligent applications, driving increasing computational demands on GPUs. However, modern GPU multitasking scheduling algorithms fail to effectively balance real-time task performance and resource utilization, especially under dynamic workloads with highly variable DNN computational demands. The complex and workload-dependent execution times of DNN kernels often lead to inefficient resource allocation, degraded system throughput, and missed real-time constraints. To address these challenges, we propose Flexible Parallel Orchestrator (FlePo), a GPU multitasking scheduling framework designed to optimize resource utilization and maintain real-time task performance within acceptable limits for soft real-time systems. FlePo integrates two key techniques: Adaptive Padding Dispatch (APD), which dynamically schedules best-effort tasks while leveraging the predictable execution characteristics of DNN kernels to maintain real-time predictability; and Dynamic Parallel Fusion (DPF), which employs kernel fusion to create computational isolation, reducing interference in parallel job execution. By combining offline profiling with online adaptation, FlePo efficiently responds to workload variations. We evaluate FlePo on two heterogeneous GPU platforms, NVIDIA Tesla V100 and AMD MI50, achieving up to a 50% increase in throughput while keeping real-time overhead below 2%. Our work enhances GPU multitasking in dynamic environments, with potential applications in autonomous driving, smart homes, and intelligent healthcare.

Förderung

  • National Natural Science Foundation of China
  • National Key Research and Development Program of China
  • Science and Technology Development Fund, Macao S.A.R (FDCT) Projects
  • General Program of China Postdoctoral Science Foundation

Zitieren

APA

Liang, H., Yang, X., Ge, R., Xia, Y., Hu, C., Zhou, X., & Cheng, D. (2026). FlePo: GPU Multitask Scheduling Optimization Framework for Dynamic Scenes. ACM Transactions on Autonomous and Adaptive Systems, 21(3), 1-35. https://doi.org/10.1145/3807775

Vancouver

Liang H, Yang X, Ge R, Xia Y, Hu C, Zhou X, et al. FlePo: GPU Multitask Scheduling Optimization Framework for Dynamic Scenes. ACM Transactions on Autonomous and Adaptive Systems. 2026;21(3):1-35. doi:10.1145/3807775

BibTeX
@article{Liang2026,
  author = {Liang, Huanghuang and Yang, Xin and Ge, Rui and Xia, Yaqi and Hu, Chuang and Zhou, Xiaobo and Cheng, Dazhao},
  title = {FlePo: GPU Multitask Scheduling Optimization Framework for Dynamic Scenes},
  journal = {ACM Transactions on Autonomous and Adaptive Systems},
  year = {2026},
  volume = {21},
  number = {3},
  pages = {1--35},
  doi = {10.1145/3807775}
}
Kostenlos starten →