Metis: Agentic Knowledge Synthesis for Explainable I/O Performance in HPC Systems
Open Access
- 13 July 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 817-826
Abstract
I/O performance explainability in HPC requires contextual characterization across the full software and system stack. Contextual characterization identifies the semantics and runtime role of I/O functions. State-of-the-art contextual characterization is still largely manual, but remains highly valuable for explaining bottlenecks and guiding optimization. However, manual contextual characterization is difficult to scale and hard to reproduce as I/O libraries and cross-layer interactions grow in complexity. We present Metis, a framework for systematic characterization of HPC I/O functions that uses agentic LLMs to integrate heterogeneous evidence sources and quantify agents agreement. Across evaluation, Metis improves held-out-category generalization over an MCP Tool baseline (0.90 vs. 0.35), reduces runtime (27.7 s vs. 84.5 s) while increasing throughput (41.8 vs. 14.28 functions/min), and sustains high verifier throughput under federated scaling (330K–1.18M functions/s). These results demonstrate that Metis is an effective and practical approach for explainable characterization of complex HDF5 behavior, enabling more trustworthy and reproducible HPC I/O analysis.Keywords
Funding Information
- LLNL LDRD (5-ERD-042)
This publication has 29 references indexed in Scilit:
- IOAgent: Democratizing Trustworthy HPC I/O Performance Diagnosis Capability via LLMs2025
- h5bench: A unified benchmark suite for evaluating HDF5 I/O performance on pre‐exascale platformsConcurrency and Computation: Practice and Experience, 2024
- Uncovering I/O demands on HPC platforms: Peeking under the hood of Santos DumontJournal of Parallel and Distributed Computing, 2023
- I/O performance analysis of machine learning workloads on leadership scale supercomputerPerformance Evaluation, 2022
- Extracting and characterizing I/O behavior of HPC workloads2022
- Parallel I/O Evaluation Techniques and Emerging HPC Workloads: A Perspective2021
- Recorder 2.0: Efficient Parallel I/O Tracing and Analysis2020
- Modular HPC I/O Characterization with Darshan2016
- A multi-level approach for understanding I/O activity in HPC applications2013
- 24/7 Characterization of petascale I/O workloads2009