EXPLORE
HPC MEETS AI
June 10 – 13, 2025
CCH, Hamburg
GERMANY
CONNECTING AI AND HPC
High performance computing (HPC) and artificial intelligence (AI) are continuing to converge, and in doing so are reshaping the HPC landscape.
This year’s ISC conference and exhibition will offer some unique opportunities for HPC-powered businesses, research organizations, and service providers to learn about the latest developments in this emerging paradigm. The event will provide a forum for exploring how AI is driving technological innovation in the field and how it is expanding the performance, energy efficiency, and scope of HPC applications.
Ready to explore the future of AI and HPC?
OFFERINGS AT ISC
Below is a synopsis of ISC 2025 sessions, workshops, and tutorials that delve into the utility of AI and how it is currently being leveraged by research organizations and businesses in the HPC community. Also provided is a summary of exhibitors offering AI technologies, products, and services.
KEYNOTES
HPC and AI - A Path Towards Sustainable Innovation
Tuesday, June 10
In the event’s opening keynote, AMD Chief Technology Officer and Executive Vice President Mark Papermaster will present his company’s vision of HPC-AI convergence and how it is driving innovation in silicon, system design, software, and energy efficiency. He will also highlight the need for an open ecosystem to facilitate this convergence.
Tackling Fragmentation in Exascale Supercomputing and Beyond
Thursday, June 12
In the closing keynote, Prof. Dr. Yutong Lu, Director of the National Supercomputing Center in Guangzhou, will talk about one of the most important challenges in the post-exascale era, specifically the fragmentation of system architectures and application requirements, brought on largely by the convergence of AI with traditional HPC. In her presentation, Prof. Lu will argue that a collaborative co-design approach involving hardware, system software, and applications will be needed to bridge these divides.
The New Landscape of Climate Computing
Wednesday, June 11
In the midweek keynote, Bjorn Stevens, Director of the Climate Physics Department of the Max Planck Institute for Meteorology, will talk about how a market for climate services can be enabled by AI and HPC. Specifically, he will describe how computationally intensive AI and HPC resources can be used to deliver local climate forecasts that can furnish actionable intelligence for governments, businesses, and research organizations.
PANELS
Trustworthiness and Energy Efficiency in AI for Science
Wednesday, June 11
This expert panel will discuss the challenges of developing reliable and energy-efficient AI for scientific applications. The exchange will address innovative hardware and software frameworks, energy-aware training models, and benchmarking metrics that will be needed to produce AI systems that are both trustworthy and practical for energy-constrained environments.
Big Questions for the Future of HPC-AI
Wednesday, June 11
This panel will explore the market shifts and other types of disruptions that are occurring due to the convergence of HPC and AI. This is a “Fishbowl Panel,” which will use a rapid-fire interactive format that includes audience participation.
Accelerating the Competitiveness of Industry through HPC and AI
Thursday, June 12
This panel of experts will discuss how combining the technologies of HPC and AI can boost competitiveness in industry. The conversation will center on three main areas: successful use cases, barriers to adoption, and the role of public-private partnerships. The issues of AI trustworthiness and sustainability will also be explored.
Birds of a feather (BOF) SESSIONS
Machine Learning for Computer-Aided Engineering
Tuesday, June 10
This BoF will explore the utility of machine learning for computational fluid dynamics (CFD) and finite-element analysis (FEA) workloads. The focus will be on the advantages of machine learning for these applications, as well as the challenges of HPC resource requirements, the general applicability of the models and their accuracy.
Artificial Intelligence and Machine Learning for HPC Workload (4th Edition)
Tuesday, June 10
The 4th iteration of this BoF will highlights AI and machine learning applications designed to analyze large-scale datasets generated by HPC workloads. This year, the discussion will focus on recent work at NERSC, NREL and BSC.
Defining Open Standards Programming for AI and HPC - Hosted by the Unified Acceleration (UXL) Foundation
Wednesday, June 11
This BoF will survey the current challenges of developing portable, open standards software across different types of AI/HPC hardware, including CPUs, GPUs, NPUs, TPUs, and other specialized processors. The session will explore the difficulties of devising new standards in the midst of ongoing development with vendor-supplied tools.
Two Worlds Collide: Trustworthiness and Sustainability for Coupled HPC and AI Simulation
Thursday, June 12
This year’s installment of the Two Worlds Collide BoF series will center on promoting a trustworthy and sustainable framework for integrating traditional HPC simulation with the rapidly emerging deep learning ecosystem. The discussion will feature representatives from Argonne National Laboratory, Oak Ridge National Laboratory, the Swiss National Supercomputing Centre, and Jülich Supercomputing Centre, as well as industry participants from AMD and Groq.
Tutorials
Friday, June 13
Please note: The tutorial pass also provides access to workshops!
Principles and Practice of Scalable and Distributed Deep Neural Networks Training and Inference
This tutorial will provide an overview of recent trends in deep learning and the role of the latest hardware and interconnects in advancing the field. Hands-on exercises are included to give attendees first-hand experience in running distributed DL training and inference on a GPU cluster.
Programming Novel AI Accelerators for Scientific Computing
This tutorial will provide an overview of the AI accelerators available at Argonne Leadership Computing Facility, including SambaNova, Cerebras, Graphcore, and Groq. Hands-on exercises are provided, focusing on the hardware features and software stacks of the accelerators.
Distributed Deep Learning on GPU-based Clusters
This tutorial will cover the basic principles of modern distributed deep learning frameworks and algorithms, using large language models (LLMs) as a running example. Students will learn to train an LLM model from scratch, as well as apply inference on the model.
WORKSHOPS
Friday, June 13
Fifth International Workshop on Computational Aspects of Deep Learning
This 5th workshop on Computational Aspects of Deep Learning (CADL) will address the challenges involved in the integration of deep learning and HPC. The workshop will explore solutions for cost-effectiveness, democratizing access to the latest research and development, and fostering AI-HPC convergence.
Trillion Parameter Consortium Workshop: Synthetic Data, Trust, and Scale When Advancing Reliable AI Models for Science
This workshop provides a forum to discuss the technical, ethical, and accessibility challenges of using synthetic data for training and evaluating AI models. A special focus will be on how to use high performance computing to generate, process, and validate synthetic data at scale.
The First International Workshop on Foundational Large Language Models Advances for HPC
This workshop will explore novel techniques to leverage LLMs for boosting the efficiency and performance of HPC tasks. Papers are being sought on various aspects of applying LLM to HPC targets, including programming environments, scientific applications, hardware design, and reliability/benchmarking metrics.
Exhibits
June 10 – June 12
This year, 70 exhibitors will be showcasing their AI products and services at ISC. Included are big names like NVIDIA, IBM, HPE, and Fujitsu, as well as more specialized providers such as Untether AI, VAST Data, and Scality.
Also in this list are HPC-powered research organizations like the Jülich Supercomputer Centre, the RIKEN Center for Computational Science, the Barcelona Supercomputing Center, GENCI, and LUMI (CSC), which have incorporated AI technology into their mission.
AI FACTORIES
A special ISC exhibitor in the R&D space is the EuroHPC Joint Undertaking (JU), an EU entity that funds efforts to develop a European supercomputing supply chain. To date, it has procured nine top-of-the-line supercomputers across the continent.
Recently, the organization was tasked with establishing AI Factories in Europe. AI Factories consist of infrastructure that use existing EuroHPC JU supercomputing capacity to develop cutting-edge AI models. The factories will serve as hubs for advancing AI throughout the EU in sectors such as health, manufacturing, climate, finance, and space, among others.
The original seven AI Factories will be located in Finland (LUMI AI Factory), Germany (HammerHAI), Greece (Pharos), Italy (IT4LIA), Luxembourg (L-AI Factory), Spain (BSC AI Factory), Sweden (MIMER). The next six selected include Germany (JAIF), France (AI2F), Poland (PIAST AIF), Austria (AIF Austria), Slovenia (SLAIF), and Bulgaria (BRAIN++).
To learn more about this remarkable investment, stop by the EuroHPC JU exhibitor booth (No. J30).