NHR4CES will once again join the NHR booth (J10) alongside other NHR centers at ISC 2025 Conference. Celebrating its 40th anniversary, ISC remains the world’s leading forum for advancing the application of high performance computing. This year’s theme „Connecting the dots“ highlights the conference’s role in fostering global collaboration, innovation, and knowledge exchange across the HPC community. The exhibition will showcase the latest developments in HPC, covering all significant advancements in system design, programming models, applications, artificial intelligence, quantum computing, and emerging technologies.
Our experts are excited to talk to interested people at the NHR booth and to meet the community. Visit NHR4CES and let’s talk about our high performance computers CLAIX-2023 and Lichtenberg II. Besides joining the NHR Alliance booth, our NHR4CES members will be actively involved in the conference program.
Our Sofa Talks
Connecting Visualization and High-Performance Computing
Time: Tuesday, June 10, 2025, 4.20 pm – 5.30 pm
Venue: NHR booth J10
Description: Join Tim Gerrits, member of our CGS Visualization, and his guests and discuss current trends, share innovative ideas, and explore potential solutions, with the goal of disseminating them across both communities
This discussion will focus on:
· The symbiotic relationship of Visualization and High-Performance Computing (HPC)
· Potential but also challenges that both the HPC and visualization communities face
· Connect researchers, developers, and users working at the intersection of HPC and visualization
Guests:
Jens Henrik Göbbert (Research Scientist for High Performance Computing and Visualization at Scale bei Forschungszentrum Jülich)
Uwe Wössner (Head of Visualization Department at the High-Performance Computing Center Stuttgart (HLRS))
François Mazen (Director – Scientific Visualization Europe @Kitware)
Jean Favre (Senior Visualisation Software Engineer @ Swiss National Supercomputing Centre (CSCS))
Marcel Krüger (Doctoroal Candidate, CSG Visualization, NHR4CES)
GPU- and ML-Benchmarking for HPC Procurements
Time: Wednesday, June 11, 2025, 1.00 pm – 2.00 pm
Venue: NHR booth J10
Description: Sandra Wienke, PI of our CSG Parallelism and Performance, hosts the Sofa Talk „GPU- and ML-Benchmarking for HPC Procurements“ join her and her guests at the NHR sofa and get an insight into GPU- and ML-Benchmarking for HPC Procurements.
As artificial intelligence (AI) and especially machine learning (ML) applications have become an integral part of operating today’s HPC (and especially GPU) clusters, HPC centers must adapt their HPC procurement processes also to meet the needs of the ML community. This includes integrating GPU- and ML-specific benchmarks into their RFPs (Request for Proposals) and acceptance tests. However, benchmarking ML-related scientific applications on HPC clusters poses (new) challenges: For example, these benchmarks need to run on different device architectures (e.g., NVIDIA, AMD GPUs) across multiple compute nodes (including GPU-aware networks), consider device memory constraints, account for reduced or mixed precision, and deliver reproducible runtimes (e.g., by defining run rules) in a given time frame. Moreover, they must be mature enough to be part of a formal HPC tender.
This sofa talk fosters the exchange of experiences in integrating GPU and ML benchmarks into HPC procurements. Participants may discuss the challenges of using different ML benchmark initiatives (e.g. MLPerf HPC, HPL-AI, Deep500, SPEC ML or LLMFoundry), different ML parts (training, inference) or data sets, and their maturity for HPC tenders as described above. Finally, this sofa talk aims to improve future HPC procurements with ML focus step by step.
Other Activities:
Lightning Talk: Picking Low Hanging Fruit of the Energy Efficiency Tree
Time: Thursday, June 12, 2025, 4.00 pm – 5.00 pm
Presenters: Christian Wassermann, Gerrit Toehgiono, Gert Vanberg, Matthias Müller, IT Center RWTH Aachen
Venue: Hall Z, 3rd floor
Half day tutorial: Advanced OpenMP: Performance and 6.0 Features
Time: Friday, June 13, 2025, 9.00 am – 1.00 pm
Venue: Hall Y12, 2nd floor
Description: Christian Terboven, PI from our CSG Parallelism and Performance, explains what participants can expect from the tutorial „Advanced OpenMP: Performance and 6.0 Features“ at ISC 2025.
Who should attend? HPC programmers with some knowledge of OpenMP that want to implement efficient shared-memory code. Intermediate Level 50%, Advanced Level 50%
What you will learn: OpenMP is the leading, portable, and widely supported shared-memory model, that developers usually find easy to learn. However, they are often disappointed with the performance and scalability of the resulting code. This disappointment stems not from shortcomings of OpenMP, but rather from the lack of depth with which it is employed.
This tutorial bridges the gap between basic usage and advanced features by exploring:
· Implications of possible OpenMP parallelization strategies, both in terms of correctness and performance
· Language features in-depth, with emphasis on advanced features like vectorization and compute acceleration
· Performance aspects, such as data and thread locality on NUMA architectures, and exploitation of the comparably new language features
· Recent additions introduced with OpenMP 6.0.
Paper Presentation: Speculative Recursion Unrolling
Time: Friday, June 13, 2025, 9.00 am – 1.00 pm
Presenters: Tim Heldmann & Sebastian Kreutzer, Scientific Computing, TU Darmstadtvier Teruel
Venue: Hall X10, 1st floor
Workshop: Energy Efficiency with Sustainable Performance: Techniques, Tools, and Best Practices
Time: Friday, June 13, 2025, 9.00 am – 2.00 pm
Presenters: Radita Liem, IT Center RWTH Aachen, Ayesha Afzal, Sarah Neuwirth
Venue: Hall X3, 1st floor
Paper presentation: Enabling Modular In-situ Workflows through CatalystMaestro and CatalystComposer
Time: Friday, June 13, 2025, 1.30 pm – 2.15 pm
Presenters: Marcel Krüger, Torsten Kuhlen, Tim Gerrits, CSG Visualization
Venue: Hall E, 2nd floor
Paper presentation: Using Transfer Learning for Predicting I/O Time Across Systems
Time: Friday, June 13, 2025, 1.30 pm – 2.15 pm
Presenters: Radita Liem, IT Center RWTH Aachen, Adrian Voß, Jay Lofstead, Julian Kunkel, Philip Carns, Mathias Müller, IT Center RWTH Aachen
Venue: Hall E, 2nd floor
