Technical Papers Archive

Constructing a Job-Centric Frontier Telemetry Dataset to Accelerate HPC Operations, Systems, and Energy Research


Authors: Leah Huk, Rachel Palumbo, Woong Shin, Tim Osborne, Ryan Adamson, Michael Sandoval, and Corwin Lester (Oak Ridge National Laboratory (ORNL))

Abstract: Comprehensive analysis of high-performance computing (HPC) systems requires linking workload execution to system behavior. This kind of analysis is vital for diagnosing performance issues, managing capacity, detecting anomalous workloads, and understanding how applications interact with system hardware. However, node telemetry data and job scheduler records are collected and analyzed in isolation and application identification requires a directed effort. We present a job-centric telemetry dataset that unifies scheduler job records with node-level measurements, enabling direct association between workloads and their corresponding power, thermal, and performance characteristics. This dataset, which spans two years of Frontier operations, aims to support a wide range of research and operational applications, including workload characterization, system analysis, and anomaly detection. This paper describes the dataset curation process, including data collection, normalization, and association of telemetry with individual job metadata. To enable further study, the dataset will be published alongside this paper for use by the HPC community.



Back to Technical Papers Archive Listing