Walking Talk Sustainability
26 seats left
Thursday, June 25, 2026 10:00 AM to 10:45 AM · 45 min. (Europe/Berlin)
Registration Counter, Entrance Hall - Ground Floor
Walking Talk
Community EngagementEnergy Efficiency and SustainabilityEnergy ManagementHW and SW Design for Scalable Machine LearningOptimizing for Energy and Performance
Information
High-performance computing centers are entering a new era of AI integration, where simulation, machine learning, and data-driven methods operate side by side to accelerate scientific discovery and expand what is computationally possible. Machine learning surrogate models are augmenting and at times superseding portions of traditional simulation workflows, significantly reducing time-to-solution and computational cost for selected problems. At the same time, this transformation is driving a shift in requirements across compute, storage, networking, data management, and facility infrastructure placing renewed focus on energy efficiency, power density, and sustainable datacenter operations, and on optimizing total energy to solution across increasingly complex workflows.
This interactive walking talk explores how HPC centers and technology partners are addressing these challenges through innovations across the full stack. Across multiple stations on the ISC exhibition floor, attendees will explore the systems and infrastructure advances enabling the next generation of sustainable accelerated computing systems. The tour will highlight how decades of HPC operational experience and innovation are shaping modern AI compute systems and datacenter infrastructure across the industry. These innovations, developed collaboratively across the ecosystem, are helping centers scale increasingly demanding simulation and AI workloads while improving efficiency and managing rising energy demands.
The tour will also explore how HPC centers and research organizations are balancing scientific capability, operational efficiency, and sustainability goals as AI and simulation workflows increasingly converge in production environments. Drawing from real-world scientific and industrial use cases, attendees will examine how organizations are managing the evolving balance between traditional simulation, AI-driven surrogate models, data curation, model training, and continuous validation workflows. By connecting these operational realities with the technologies showcased on the exhibition floor, the session provides a practical perspective on the computational and energy tradeoffs shaping the next generation of sustainable HPC and AI infrastructure.
This interactive walking talk explores how HPC centers and technology partners are addressing these challenges through innovations across the full stack. Across multiple stations on the ISC exhibition floor, attendees will explore the systems and infrastructure advances enabling the next generation of sustainable accelerated computing systems. The tour will highlight how decades of HPC operational experience and innovation are shaping modern AI compute systems and datacenter infrastructure across the industry. These innovations, developed collaboratively across the ecosystem, are helping centers scale increasingly demanding simulation and AI workloads while improving efficiency and managing rising energy demands.
The tour will also explore how HPC centers and research organizations are balancing scientific capability, operational efficiency, and sustainability goals as AI and simulation workflows increasingly converge in production environments. Drawing from real-world scientific and industrial use cases, attendees will examine how organizations are managing the evolving balance between traditional simulation, AI-driven surrogate models, data curation, model training, and continuous validation workflows. By connecting these operational realities with the technologies showcased on the exhibition floor, the session provides a practical perspective on the computational and energy tradeoffs shaping the next generation of sustainable HPC and AI infrastructure.
Format
on-site
