slideshow 1

You are here

Technische Universität München (TUM)

Organization

Technische Universität München is one of the Excellence Universities of Germany. The Faculty of Informatics is one of the 12 faculties of the university. It has 20 chairs, over 30 professors, and about 1700 bachelor and master students. The chair on Computer Architecture (LRR) has been working in the field of parallel processing for over 20 years. It currently comprises two professorships and about 20 scientists.

Expertise

LRR‘s major research is in interactive run-time tools for parallel and distributed systems. LRR has developed performance analyzers, debuggers, program flow visualizers, and the necessary on-line monitoring techniques. In addition, LRR investigates high-performance cluster computing as well as accelerated HPC architectures. It participated in several ESPRIT and IST projects related to tools and distributed computing. One of the focus areas at LRR is the development of Periscope an automatic online performance analysis framework. Within several projects Periscope became a production-level tool which is used on highly scalable HPC systems for performance analysis of large scale applications. The main funding is currently from two projects SILC and ISAR from the German Ministry for Education and Research (BMBF). Both projects are collaborations with industry and large supercomputing facilities in Germany.

Role

TUM is the main developer of Periscope. It will provide its extensive knowledge on performance analysis and tuning for scientific parallel applications as well as its deep knowledge on the design and implementation of Periscope. It will also bring in its extensive collaboration with other European and American groups on performance analysis tools within the Vi-HPS as well as beyond this virtual institute.

Key personnel

Michael Gerndt received a Ph.D. in Computer Science in 1989 from the University of Bonn. He developed
SUPERB the first automatic parallelizer for distributed memory parallel machines. For two years, in 1990 and 1991, he held a postdoc position at the University of Vienna and joined Research Centre Jülich in 1992 where he concentrated on programming and implementation issues of shared virtual memory systems. This research led to his habilitation in 1998 at TU München. Since 2000 he is professor for architecture of parallel and distributed systems at TU München. Michael Gerndt was the coordinator of the FP6 European American Working Group APART (Automatic Performance Analysis: Real Tools).

His current research focuses on programming models and tools for scalable parallel architectures. He is leading the development of the automatic performance analysis tools Periscope and of iOMP, an extension of OpenMP for invasive computing. iOMP is a research project in the new Transregional Collaborative Research Center InvasIC (TR 89) funded by the German Science Foundation. In addition he is heading projects on parallel programming languages and their implementation on multicore processors as well as resource management in Cloud environments funded by public and industry sources. Michael Gerndt is the contact person of the Faculty of Informatics for international affairs.

Result dissemination and exploitation

TUM has been developing online performance analysis tools since many years. This project will start an extension of this work towards online tuning which will be of enormous importance for future programming of parallel architectures from the embedded area up to the high end. The results of this project will be published in conferences and journals. They will also be integrated into the lectures at TUM. The online tuning framework will be available to application people from the HPC community for download. It will not only be advertized by TUM but will become part of the tools suite of the VI-HPS and thus part of the tuning workshops held at supercomputing centers of the Gauss Allianz and at European and worldwide supercomputing centers. These courses combine presentations and hands-on sessions with codes of application groups to familiarize those groups with the tools provided by VI-HPS. In addition to the tuning workshops TUM collaborates with the other groups in VI-HPS in the presentation of performance analysis tutorials at major parallel computing conferences.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer