Division Application Support
Head of the division Application Support
|Dr. Norbert Attig|
Phone: +49 2461 61-4416
Fax: +49 2461 61-6656
The mission of this division is the support of HPC users and the efficient usage of their applications on the HPC systems at Forschungszentrum Jülich. Our users are based at German universities, national research institutes, and the Research Centre itself; increasingly they come from European institutes outside of Germany, from industry, and from European infrastructure projects like PRACE, LinkSCEEM, ESMI, and SOPHIA, in which our institute is actively involved. In recent years the European part of our user community has been steadily growing.
To accomplish this mission, the division is structured in three groups:
High-level User Support
The focus of this group is to provide expert support in high-performance computing (HPC) for users of the supercomputing facilities at Forschungszentrum Jülich. Furthermore, this group collaborates in the implementation of the PRACE and LinkSCEEM infrastructures.
- Support and service: As support we advice our users in efficient use of our systems. Additionally, as cross-sectional team "Application Optimization" codes are ported and optimized together with developers and users. Extensively used software packages are provided, maintained, and benchmarked on Jülich's HPC systems.
- Education and training: Courses on MPI and OpenMP as well as parallel I/O are provided, introductory courses about our HPC systems are organized, and the group takes part in the PRACE Advanced Training Centres (PATCs) programme. Our contribution to education comprises the supervision of bachelor and master thesis.
- Development of tools focussing on application monitoring and benchmarking for HPC users: LLview, Jumel, and JuBE. Furthermore, the development of synthetic benchmarks is part of our work.
Research is focussed on high-performance I/O and resulted in the design of the widely used library SIONlib. Further research on I/O is performed in collaboration with the Exascale Innovation Center (EIC).
Programming Environments and Performance Analysis
The HPC SoftWare Analysis and Tools Team contributes in the four areas:
- Service: Optimization user application codes as cross-sectional team Performance analysis in collaboration with the Simulation Laboratories and cross-sectional teams Application Optimization and Methods and Algorithms.
- Education: Training of HPC users in the area of performance analysis and tools as part of PRACE PATC and VI-HPS training activities, summer and winter schools as well as supervision of bachelor, master and Ph.D. thesis in cooperation with RWTH and GRS.
- Development: Design, implementation and maintenance of performance tools around the Scalasca toolset
- Research: Participation in national and international research projects and the Exascale efforts IESP, EESI, EIC, ECL.
Multimedia, Visualization, Virtual Reality
The goal of the visualization group is the visualization of scientific results with state of the art algorithms and techniques.
- Support and service: Hard- and Software systems are provided for the visualization of scientific data with the aim to analyze and present scientific results. This includes visualization nodes integrated in our JURECA HPC system, stereoscopic output devices and equipment for video editing and multimedia production. Together with the HPC users, interactive visualization applications, videos and multimedia contents are produced.
- Education and training: Lectures on "Algorithms and Data Structures" and on "Scientific Visualization" for the Bachelor's course MATSE and the Master's course Technomathematics are given in cooperation with the Aachen University of Applied Science, campus Jülich. Training courses addressing scientific visualization with the software packages OpenGL, IDL, VTK, Paraview and Visit are offered to the HPC users.
Development and research: In cooperation with the topical scientists, the cross-sectional team Visualization implements new algorithms and applications for data visualization for desktop as well as for large screen 3D visualization systems. Interactive virtual reality visualization techniques are developed in collaboration with the RWTH Aachen within the Jülich Aachen Research Alliance JARA, section JARA HPC.