A Real-Time Digital Twin for Adaptive Scheduling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Yihe, Kurkure, Yash, Tao, Yiheng, Papka, Michael E., Lan, Zhiling
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917162644406272
author Zhang, Yihe
Kurkure, Yash
Tao, Yiheng
Papka, Michael E.
Lan, Zhiling
author_facet Zhang, Yihe
Kurkure, Yash
Tao, Yiheng
Papka, Michael E.
Lan, Zhiling
contents High-performance computing (HPC) workloads are becoming increasingly diverse, exhibiting wide variability in job characteristics, yet cluster scheduling has long relied on static, heuristic-based policies. In this work we present SchedTwin, a real-time digital twin designed to adaptively guide scheduling decisions using predictive simulation. SchedTwin periodically ingests runtime events from the physical scheduler, performs rapid what-if evaluations of multiple policies using a high-fidelity discrete-event simulator, and dynamically selects the one satisfying the administrator configured optimization goal. We implement SchedTwin as an open-source software and integrate it with the production PBS scheduler. Preliminary results show that SchedTwin consistently outperforms widely used static scheduling policies, while maintaining low overhead (a few seconds per scheduling cycle). These results demonstrate that real-time digital twins offer a practical and effective path toward adaptive HPC scheduling.
format Preprint
id arxiv_https___arxiv_org_abs_2512_18894
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Real-Time Digital Twin for Adaptive Scheduling
Zhang, Yihe
Kurkure, Yash
Tao, Yiheng
Papka, Michael E.
Lan, Zhiling
Distributed, Parallel, and Cluster Computing
High-performance computing (HPC) workloads are becoming increasingly diverse, exhibiting wide variability in job characteristics, yet cluster scheduling has long relied on static, heuristic-based policies. In this work we present SchedTwin, a real-time digital twin designed to adaptively guide scheduling decisions using predictive simulation. SchedTwin periodically ingests runtime events from the physical scheduler, performs rapid what-if evaluations of multiple policies using a high-fidelity discrete-event simulator, and dynamically selects the one satisfying the administrator configured optimization goal. We implement SchedTwin as an open-source software and integrate it with the production PBS scheduler. Preliminary results show that SchedTwin consistently outperforms widely used static scheduling policies, while maintaining low overhead (a few seconds per scheduling cycle). These results demonstrate that real-time digital twins offer a practical and effective path toward adaptive HPC scheduling.
title A Real-Time Digital Twin for Adaptive Scheduling
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2512.18894