HPC-ODA 2026 - 1st International Workshop on HPC Operational Data Analytics

Held in conjunction with IEEE/ACM SC 2026, The International Conference for High Performance Computing, Networking, Storage and Analysis, in Chicago, IL

Workshop website: https://hpc-oda.org/workshop2026


Description
===========
High-performance computing (HPC) systems generate vast streams of operational data: system logs, job scheduler records, hardware metrics, energy consumption, network performance, and environmental telemetry.

This data is foundational to the entire HPC ecosystem. Operations teams rely on it for system reliability, efficiency, and predictable maintenance. Researchers need it for training AI/ML models, constructing digital twins, and studying system behavior at scale. Users need it to understand application behavior, diagnose performance bottlenecks, and optimize resource utilization.

Yet access to this data — and the ability to analyze it effectively — remains unevenly distributed across these communities. The HPC Operational Data Analytics (HPC-ODA) Workshop is a dedicated venue for bridging that gap.

Building on eight years of successful Birds-of-a-Feather sessions at SC and ISC, this inaugural workshop transitions to a peer-reviewed format to enable deeper technical engagement and lasting community collaboration. The workshop welcomes contributions from researchers, practitioners, operators, tool developers, students, and others working at the intersection of HPC systems, telemetry, monitoring, and analytics.


Topics of Interest
==================
We welcome submissions on any aspect of operational data analytics in HPC, including but not limited to:

* ODA Infrastructure and Engineering
o System integration and architecture
o Scalable data collection, ingestion, and processing pipelines
* Data Visualization and Analytics
o Analysis methods and techniques
o Machine learning applications
o Real-time streaming analytics
* Data Standardization, Management, and Governance
o Data lifecycle management
o Security and compliance
o FAIR data practices
* HPC Use Cases and Data Impact
o Energy and sustainability
o Application runtime profiling
o Performance engineering
o Anomaly detection
o HPC lifecycle and procurement
o Business operations and ROI
* Best Practices and Case Studies
o Implementation experiences
o Deployment strategies
o Performance validation
* Emerging Technologies
o Digital twins
o Large language models applied to HPC operations
o Multi-site and cloud-HPC hybrid operations
o Advanced visualization


Submissions
===========

Papers
------
Papers may be submitted as either:
* Short papers — 4 two-column pages, appropriate for work-in-progress or focused contributions
* Full papers — 8 two-column pages, for mature research contributions

All papers must be submitted in PDF format using the IEEE conference template via the SC Linklings submission system. Each submission will receive a minimum of three single-blind peer reviews.

In support of the SC26 Reproducibility Initiative, authors may optionally submit an Artifact Description and Artifact Evaluation (AD/AE) appendix. This appendix will not be reviewed but will be included with papers after acceptance, in accordance with SC26 guidelines.

Accepted papers will be included in the workshop program and presented during the event. Presentation format and duration will be determined based on the final program.


Lightning Talks
---------------
Lightning talks provide an opportunity to present emerging work, operational experiences, tools, dashboards, and other contributions that may not be suitable for full or short paper submissions. Submissions are welcome from across the HPC community, including researchers, practitioners, and students. We especially encourage contributions that share early-stage ideas, operational insights, or practical experiences.

Lightning talk submissions should consist of an extended abstract (1–2 pages). There is no strict formatting requirement; however, submissions should be clear, well-structured, and comparable in scope to a short paper (e.g., including motivation, approach, and key outcomes). Submissions should be provided as a PDF.

Submissions will be peer-reviewed for relevance and quality. Accepted lightning talks will be allocated a 5-minute presentation slot and will be documented in the HPC-ODA workshop report published through the EE HPC WG.


AI Use Policy
-------------
Content generated by AI tools must be clearly and thoroughly described in the submission. AI systems may not be credited as authors.


Important Dates
===============
All deadlines are 11:59 PM AoE (Anywhere on Earth).

* July 31, 2026 — Full / Short Paper Submission Deadline
* September 4, 2026 — Full / Short Paper Notification
* September 18, 2026 — Camera-Ready Deadline
* September 18, 2026 — Lightning Talk Submission Deadline
* September 25, 2026 — Lightning Talk Notification


Publication
===========
Papers accepted and presented at HPC-ODA 2026 will be published as part of the SC26 workshop proceedings in the IEEE Xplore Digital Library.

A workshop report covering the keynote, lightning talks, and open discussion will be published and distributed through the EE HPC WG Operational Data Analytics team website.


Organizing Committee
====================
* Michael Ott, LRZ
* Ayse Coskun, Boston University
* Jeff Hanson, HPE
* Melissa Romanus, NERSC/LBNL
* Woong Shin, ORNL
* Tim Osborne, ORNL