Authors: Haritha Siddabathuni Som (Argonne National Laboratory), Rebecca Hartman-Baker (Lawrence Berkeley National Laboratory (LBNL)), Chris Fuson (Oak Ridge National Laboratory (ORNL)), Smita Darmora (University of Illinois Chicago), Thomas Breuer (Jülich Supercomputing Centre (JSC)), Helena Vela Beltran (Do IT Now), Ilya Zhukov (Jülich Supercomputing Centre (JSC))
Abstract: As HPC enters an era of extreme heterogeneity and AI integration, user support is a critical strategic enabler. However, rising system complexity and finite staffing strain traditional models. To thrive in the AI and exascale era, a collaborative community approach is essential.
This BoF session invites practitioners, providers, and users to collaboratively reshape HPC support. Featuring insights from facility leaders on managing rapid AI adoption, an interactive live survey, and open group discussions, participants will share experiences regarding scaling, AI automation, and knowledge transfer. Outcomes will form a post-conference report dedicated to fostering a sustainable, community-driven support ecosystem.
Long Description: Motivation and Scope
HPC is entering a new phase characterized by extreme heterogeneity and tight
integration with AI workflows, fundamentally transforming user interactions and
increasing demands on user support. Across research and industry, support teams face
critical bottlenecks as they serve larger, more diverse communities and navigate rapidly
evolving AI frameworks under tight staffing constraints. Despite its direct impact on
system utilization and scientific output, HPC user support remains local, fragmented, and
weakly coordinated. This BoF brings together practitioners and stakeholders to examine
how support must evolve to remain sustainable in the AI and exascale era.
Relevance to the SC Community
SC brings together practitioners, infrastructure providers, researchers, and industry
partners who rely on small- and large-scale HPC systems. While system architecture and
software performance are core SC topics, user support, training, and knowledge transfer
are equally critical to achieving high utilization, return on investment, and scientific
output. By treating user support as a strategic enabler rather than operational overhead,
this BoF directly aligns with SC’s goals of collaboration, knowledge exchange, and
building sustainable, cross-institutional HPC ecosystems.
Focus
The BoF centers on a core question: How can HPC user support scale sustainably and
effectively in an environment shaped by AI, system heterogeneity, and growing user
diversity? Starting with the session introduction, followed by a brief 10-minute
introduction of the discussion topics, we will dive into parallel group discussions to
address key topics, some of them could be:
Scaling user support under increasing demand and complexity introduced by AI
Tooling, automation, and the role of AI in user support
Challenges in adopting AI practices, including knowledge transfer, training, and
skill development
Managing user expectations and defining support quality
Opportunities for cross-centre collaboration and shared practices
Recognition and visibility of HPC user support work
Target Audience
Participation is open to all attendees involved in HPC user support across research and
industry. HPC users are highly encouraged to attend; their input is invaluable, and this
session offers a unique opportunity to connect directly with the support staff and system
administrators behind ticket systems and hotlines.
Session Format
The BoF combines short opening remarks for each topic with interactive group
discussions, minimizing long monologue presentations in favor of targeted insights, peer
exchange, and collective problem-solving. A classroom setup is requested to comfortably
accommodate both presentation viewing and small-group interaction.
Materials and Technical Setup
Each group will feature a structured note template (challenges, practices, ideas) and
markers. We will use a survey/poll software like Mentimeter to help keep the session
engaging and interactive.
Outcomes and Follow-up
Delivers actionable outputs and identifies concrete ideas for pilot exploration. Documents
insights, lessons learned, and promising practices via an open white paper or blog post.
Encourages organizers with operational responsibility to attempt pilot implementations
where feasible.
Additional Note
A similar BoF was conducted at the ISC26 conference with great success
(https://epicurehpc.eu/2026/07/09/rethinking-hpc-user-support-for-ai-and-exascale-era/
(https://epicurehpc.eu/2026/07/09/rethinking-hpc-user-support-for-ai-and-exascale-era/)). For the SC conference we aim to target a more global group to enrich the discussions.