Keynote Abstracts
Joint Keynote I: From JUPITER to the Stars – Experience and Results from Europe’s First Exascale System
– Andreas Herten (Jülich Supercomputing Centre)
JUPITER is Europe’s first Exascale system, utilizing about 24 000 GPUs to reach 1 ExaFLOP/s in the HPL benchmark. The system is hosted by Jülich Supercomputing Centre in Germany and currently finishing build-up procedures. The talk will show details of the system, its procurement and deployment, and also first results gained with these thousands of GPUs.
Andreas Herten: Andreas Herten is a lead of the division “Novel System Architecture Design” at Jülich Supercomputing Centre, where his research focuses on accelerator-based computing. He and his team optimize applications, benchmark hardware, develop tools, and performance-critical libraries. He is part of the JUPITER core team and responsible for the benchmarking activities.
Joint Keynote II: Who Decides? Beyond Runtime Defaults, Toward Adaptive Multi-Device OpenMP
– Florina M. Ciorba (University of Basel)
High performance computing (HPC) has long driven breakthroughs in physics, chemistry, and engineering. Today, the emergence of digital twins in healthcare introduces a new frontier: personalized, physics-informed simulations of the human vascular system. These models demand solving fluid dynamics over complex 3D anatomies across millions of heartbeats, while integrating continuous data from wearable sensors. The result is petabyte-scale datasets and real-time simulation needs that stretch the limits of algorithms, data handling, and scalability. This keynote will highlight how vascular digital twins expose new challenges and opportunities for HPC—reducing communication overhead in parallel time integration, compressing multimodal data streams without losing fidelity, and enabling adaptive, continuous simulation at exascale. Meeting these challenges requires leadership-scale systems co-designed with novel algorithms and workflows. Beyond medicine, these lessons illustrate how HPC can evolve to support time-critical, data-rich applications across domains, underscoring the need for sustained investment and long-term vision in high performance computing.
Florina M. Ciorba: Florina M. Ciorba is Professor of High Performance Computing and head of the HPC Lab at the University of Basel, Switzerland, which she founded in 2015. She earned her Ph.D. from the National Technical University of Athens, Greece in 2008, followed by postdoctoral positions at Mississippi State University, USA, and Dresden University of Technology, Germany. Her research focuses on scheduling and resource management, and on enhancing performance, resilience, portability, reproducibility, and autonomy in HPC and AI/ML systems. Recent work includes autotuning, energy-efficient cosmological simulations, and autonomy loops for system performance. Her group developed LB4OMP and Auto4OMP, open-source extensions of the LLVM OpenMP runtime for dynamic loop scheduling and automated scheduling algorithm selection. She has published nearly 100 peer-reviewed papers and received several best paper awards. She is a founding board member and PI of the Basel node in the SKACH project (Swiss SKAO Consortium) and co-founded IDEAS4HPC, the Swiss chapter of Women in HPC. She also serves on other various boards and committees related to HPC and energy efficiency. She is a senior & life member of ACM and active in IEEE, HiPEAC, and DISCOVER-US. More at http://hpc.dmi.unibas.ch/
Joint Keynote III: Who Decides? Beyond Runtime Defaults, Toward Adaptive Multi-Device OpenMP
– Jeff Hammond (NVIDIA)
Agentic development has progressed rapidly over the past two years, to the point where it makes sense to ask whether AI can write complex communication software by itself. In this talk, I will describe my experience building an MPI library from scratch using AI. With expert guidance on design and testing, AI was able to implement all of MPI-5 from scratch in less than a month, with support for shared-memory, sockets, OFI/libfabric and UCX, with optimized algorithms for message matching, collectives, etc. I will also talk about agentic development of GPU communication software based on NCCL and NVSHMEM, demonstrating that AI is not limited to CPU environments.
Jeff Hammond: Jeff is is a Distinguished Engineer at NVIDIA in the data center software organization, focused on GPU communications (NCCL and NVSHMEM). He has extensive experience with the design and use of parallel programming models and scientific applications. Jeff’s most notable achievements include the MPI-5 Application Binary Interface standard, development of the MPI-3 one-sided communication software ecosystem, and contributions to the NWChem quantum chemistry project. He received a PhD in Chemistry from the University of Chicago in 2009.