2023-06-12 · Leipzig, Germany
Times: CEST
Gunther Weber Rethinking High Performance Data Analysis for the Exascale and Beyond Transitioning towards exascale computing exacerbates existing challenges in supercomputing, such as the widening gap between compute power and I/O bandwidth, as well as the need to move to even higher levels of parallel processing. Limitations imposed by I/O bandwidth creates a growing necessity for adopting in situ analysis techniques. Topological data analysis provides a versatile framework to define high-level features across a broad spectrum of simulations and to automate the selection of visualization parameters such as isovalues for isosurface extraction. The global nature of topological methods hinders their parallelization, posing additional difficulties. A fresh perspective on topological data analysis enables the development of novel algorithms that effectively leverage exascale machines. Looking forward and beyond simulations, the need to analyze the ever-increasing volume of experimental and observational provides additional challenges and opportunities for high-performance computing. |
Gunther H. Weber received a Ph.D. in computer science, with a focus on computer graphics and visualization, from the University of Kaiserslautern, Germany in 2003. He is currently a Staff Scientist in the Computational Research Division at the Lawrence Berkeley National Laboratory (LBNL), where he serves as Deputy Group Lead of the Machine Learning and Analytics Group in the Scientific Data Division. Gunther Weber is also an Adjunct Associate Professor of Computer Science at the University of California, Davis. His research interests include computer graphics, scientific visualization, data analysis with using topological methods, parallel and distributed computing for visualization and data analysis applications, hierarchical data representation methods, and bioinformatics. He has extensive experience in working with researchers from diverse science and engineering fields, including applied numerical computing, combustion simulation, gene expression, medicine, civil engineering, cosmology, climate and particle accelerator modeling. Dr. Weber has authored or co-authored over 100 publications, six of which won best paper awards. He has served as principal investigator (PI) or Co-PI on several Department of Energy (DOE) and National Science Foundation (NSF) projects. He is a reviewer for major funding agencies (DOE, NSF), conference proceedings and journals. Dr. Weber served as co-organizer, co-chair and program committee member of more than 60 internationally recognized conferences.