Abstract
Hardware performance counters (HPCs) enable the measurement of microarchitectural events, which are crucial for tracking and predicting program behavior. High-fidelity measurement and precise attribution are essential for accurate profiling. However, existing profiling tools have fundamental challenges in both aspects. In measurement, numerous events compete for limited hardware monitoring resources; while for attribution, applications have diverse requirements, but systems provide limited support. Existing tools mitigate the former limitation through event multiplexing, but this approach introduces non-trivial errors. The latter limitation, however, remains largely unaddressed. This paper introduces Tintin, an HPC profiling infrastructure with a modular three-component design that addresses both challenges. Tintin introduces mechanisms to mitigate multiplexing errors by characterizing uncertainty at runtime, scheduling events to minimize it, and reporting uncertainty to applications. It also proposes the Event Profiling Context (ePX) as a new OS primitive to unify diverse profiling requirements. Tintin is evaluated using benchmarks as well as real-world resource orchestration, performance debugging, and intrusion detection systems, to demonstrate its ability to improve hardware profiling with low runtime overhead.
Recommended Citation
A. Li et al., "Tintin: A Unified Hardware Performance Profiling Infrastructure To Uncover And Manage Uncertainty," Proceedings of the 19th Usenix Symposium on Operating Systems Design and Implementation Osdi 2025, pp. 575 - 593, USENIX Association, Jan 2025.
The definitive version is available at https://doi.org/https://www.usenix.org/system/files/osdi25-li.pdf
Meeting Name
19th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2025)
Department(s)
Computer Science
Publication Status
Free Access
International Standard Book Number (ISBN)
978-193913347-2
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2026 The Authors, All rights reserved.
Publication Date
01 Jan 2025
Included in
Databases and Information Systems Commons, Other Computer Sciences Commons, Systems Architecture Commons
Comments
Acknowledgements: We thank our shepherd,Pedro Fonseca, and the anonymous reviewers for their valuable feedback. We also thank Tobias Pristupin and Lien Zhu for their contributions tothe implementation.
This work was supported by the (CNS-2154930, CNS-2229427,CNS-2141256,CNS2403758,CNS-2229290), the ARO(W911NF-24-1-0155), the ONR(N00014-24-1-2663),a WashU OVCR seed grant, and Intel.
Recommended Citation: Li, A., Sudvarg, M., Li, Z., Baruah, S., Gill, C., & Zhang, N. Tintin: A Unified Hardware Performance Profiling Infrastructure to Uncover and Manage Uncertainty. In Proceedings of the 19th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2025). USENIX Association, 2025.