Workshops Archive

CQTwin: A Digital-Twin Toolkit for Operational Analytics of HPC Schedulers


Workshop: Workshop on Operational Data Analytics in HPC (HPC-ODA)

Authors: Kanglin Xu, Yash Kurkure, and Yihe Zhang (University of Illinois Chicago); Michael Papka (Argonne National Laboratory, University of Illinois at Chicago); and Zhiling Lan (University of Illinois Chicago, Argonne National Laboratory)

Abstract: HPC facilities collect years of job scheduler records, yet the data remains underused because accessible tools for turning it into insight are scarce. We present CQTwin, an open-source digital-twin toolkit for operational analytics of HPC scheduling data. Built on the Python-based CQSim discrete-event simulator, CQTwin combines a fidelity-validated, high-performance C++ engine with an interactive browser-based dashboard and a Snapshot Lab for what-if exploration. As an offline twin, CQTwin replays archived workload traces under alternative policies, revealing how the same workload would have unfolded under different decisions. As an online twin, it consumes job events as they arrive and projects how alternative policies would process the currently queued jobs, supporting real-time decision-making. The dashboard covers the complete workflow---workload characterization, experiment configuration, and multidimensional comparison---letting users see which policies suit a workload and at what trade-offs. We validate CQTwin on Argonne Leadership Computing Facility production traces spanning 2016--2023.


Back to Workshop on Operational Data Analytics in HPC (HPC-ODA) Archive Listing

Back to Full Workshop Archive Listing