Workshop on Software Frameworks for Scalable Scientific Simulations at ISC High Performance 2015
The workshop is planned to start a series of workshops in which the advances in techniques, necessary tools, and realization of end-to-end parallel simulation frameworks are discussed. These discussions will be triggered and inspired by accepted talks of the workshop. This first installment of the workshop will provide a chance for developers from academia and industry as well as users alike to come together and talk and hear about the advances in the field of developing simulation frameworks in the HPC ecosystem. An emphasis will be put on the real-world use-cases that are solved with the presented methods, with a focus on the scalable parallelism in all necessary parts to obtain results from large scale simulations on massively parallel systems.
The keynote will be held by Carsten Burstedde, Professor for Scientific Computing and head of a research group at the University of Bonn.
Online Workshop DocumentTo track the discussions of the workshop we set up a Google Doc for collaborative editing.
- José Gracia, HLRS
- Daniel Harlacher, STS
- Harald Klimach, STS
- Sabine Roller, STS
- Martin Schulz, CASC/LLNL
Dr. José Gracia originally has a background in astrophysics and obtained his Diploma and Doctorate degree from the University of Heidelberg in 1998 and 2002, respectively. In the following years he was recipient of two European Union Marie Curie Fellowships in Greece and Ireland. Later on, he became interested in higher-level abstractions of parallel computing, focusing particularly on developer productivity. He joined the High Performance Computing Center Stuttgart (HLRS) in 2009. His research group does research on topics as parallel debugging and correctness checks, new concepts for parallel programming models at scale, and fault-tolerance for MPI. In the recent past José has (co-) organized several editions of the International Parallel Tools Workshop (2014, 2012, 2011, 2010) and the PROPER workshop (at Euro-Par 2014), has been area co-chair for INFOCOMP 2013 & 2014, and PC member of ExaMPI (at SC13), HLPGPU (at HiPEAC 2014), HetAr (at ParCo 2013), MHPC (at EuroPar 2013), and ICPADS 2013.
Dipl.-Ing. Daniel F. Harlacher originally has a background in aerospace engineering and obtained his Diploma from the University of Stuttgart in 2009. In 2009 he joined the German Research School for Simulation Sciences GmbH (GRS) in particular the department for Applied Supercomputing in Engineering (ASE). In his research he focuses on the development and the application of highly scalable simulation software for technical applications. Since 2013 he works in the STS group at the University of Siegen. From 2010 on he teaches biannually in the context of the parallel training courses offered by the HLRS.
Harald Klimach studied aerospace engineering at the University of Stuttgart with a thesis on coupled flow simulations. He worked at the national super-computing center in Stuttgart (HLRS) in the european HPC projects DEISA and PRACE. Besides his research on coupled simulations on heterogeneous systems he supported users and ported and optimized various codes from different fields for various supercomputing systems in Europe. From 2009, he worked at the German Reseach School for Simulation Sciences GmbH in Aachen and is the main architect designing and developing the end-to-end parallel APES Framework. Since 2013 he works in the STS group at the University of Siegen.
Sabine Roller is Professor for Simulation Techniques and Scientific Computing and Director of the Center for Information and Media Technology (ZIMT) at University of Siegen, Germany. She studied Technical Mathematics at University Karlsruhe and received her PhD from the University of Stuttgart. After leading the Scalable Computing group at High Performance Computing Center Stuttgart (HLRS), she became full professor in Applied Supercomputing in Engineering at RWTH Aachen University and German Research School for Simulation Sciences GmbH. Her research focuses on the simulation of large scale technical systems based on uid dynamics with multi-physics and multi-scale problems, and a key aspect is their efficient implementation on current and future supercomputing systems. Prof. Roller won an Amelia Earhart Fellowship Award in 1998/1999.
Martin is a Computer Scientist at the Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory (LLNL). He earned his Doctorate in Computer Science in 2001 from the Technische Universität München (Munich, Germany). He also holds a Master of Science in Computer Science from the University of Illinois at Urbana Champaign. After completing his graduate studies and a postdoctoral appointment in Munich, he worked for two years as a Research Associate at Cornell University, before joining LLNL in 2004. Martin is a member of LLNL's Scalability Team, which focuses on research towards a scalable software stack for next generation systems, as well as LLNL's ASC CSSE ADEPT (Application Development Environment and Performance Team) and he works closely with colleagues in CASC's Computer Science Group (CSG) and in the Development Environment Group (DEG).
- Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory (LLNL)
- High Performance Computing Center Stuttgart (HLRS)
- Simulation Techniques and Scientific Computing (STS)
The Center for Applied Scientific Computing (CASC) conducts world class, collaborative scientific research and development in mathematics, computer science, and data science on problems critical to national security. This includes the development of well known libraries like the hypre solver library or the SAMRAI adaptive mesh refinement system. Further, LLNLs Scalability Team, which is part of CASC, focuses on research towards a scalable software stack for next generation systems, incl. performance and debugging tools, performance and scalability optimization, resilience, power aware computing and parallel programming models.
HLRS is the first German national high-performance computing center. It provides supercomputing resources to academic and industrial user from all areas. Historically, its user community has a strong focus on engineering applications. HLRS draws on a long-standing expertise in parallel computing and actively in parallel software development as well as training in these areas.
The STS group at the University of Siegen develops an integrated parallel simulation framework for engineering applications. This software stack is used by the group and others on various HPC systems. The team also actively teaches in training for HPC, like the parallel training courses offered by the HLRS, the PRACE code porting workshop or HPC-Europa.
Welcome & Introduction
02:00 pm - 02:15 pm
Recent Developments in forest-of-octrees AMR
02:15 pm - 02:45 pm
Carsten Burstedde, University of Bonn
The Uintah Software Framework for Scalable Scientific & Engineering Simulations
02:45 pm - 03:05 pm
Martin Berzins, University of Utah
03:05 pm - 03:55 pm
03:55 pm - 04:00 pm
04:00 pm - 04:30 pm
Challenges in the Design of Scalable Algorithms for Shape Optimization
04:30 pm - 04:50 pm
Martin Siebenborn, Trier University
Reduced Representations for In-Situ Visualization
04:50 pm - 05:10 pm
Steffen Frey, University of Stuttgart
05:10 pm - 05:50 pm
Summary & Further Steps
05:50 pm - 06:00 pm