Research Center High Performance Computing fz-hpc-logo

Research Objectives

High Performance Computing has undoubtedly become a key technology for science, industry, and the economy over the past 10 years. New applications that benefit from effective parallelization are constantly being created. High-performance computing is a driving force in the development of new forms of medicine, in the search for renewable, and clean energy sources, in the implementation of efficient and intelligent transport systems, in climate modeling and natural disaster prediction and management, in optimizing food production as part of a more sustainable agriculture, in the development of effective security mechanisms for IT infrastructures, etc. Further growth of HPC is to be expected and indispensable for competitive research and development.

read more » 

The University of Innsbruck uses HPC primarily for computer simulations ("in silicio" experiments) in order to answer domain science questions that otherwise could not be answered or only with great effort via direct observation and measurement or real ("in vitro") experiments. In addition to theory and experiments, simulation has thus established itself as the third pillar of science. Simulation programs are mapped to powerful parallel computers using parallelization and optimization techniques as well as numerical algorithms. Of particular interest are data- and computation-intensive applications that can only be executed on massively parallel high-performance computers for realistic problem sizes within a reasonable amount of time.

However, much of the performance enhancement in high-performance computing is no longer achieved today through the evolution of hardware alone, but primarily through intelligent and customized software and application development. The FZ HPC of the University of Innsbruck organizes and supports various activities for the dissemination and efficient use of high-performance computing at the University of Innsbruck.

1. Support for the optimization and parallelization of HPC applications for supercomputers

Advanced application developers from various scientific disciplines will be supported in the parallelization and optimization as well as in the selection of numerical methods. The focus is on applications with realistic chances to be ported on the fastest supercomputers in Europe (e.g. in the context of the PRACE^or EuroHPC programme).

2. Research collaborations

The research center should also act as a partner for project proposals in the area of HPC both at national (e.g. FWF or FFG) and at international level (e.g. FWF DACH, EU H2020, EU FP9) either as an intermediary (e.g. PRACE/EuroHPC applications) or as a participant in the project. Furthermore, we seek to establish collaborations with EU centers of excellence (e.g. POP, EoCoE, E-CAM, COEGSS), whose objective is to achieve sustainable scientific HPC applications and industrial competitiveness.

3. Training in programming, performance analysis and debugging for parallel computers

For beginners as well as for advanced programmers and application developers training programs in the form of hands-on tutorials for the most important programming concepts and languages as well as performance analysis and debugging are offered. The training courses are organized at the University of Innsbruck together with the participating organizational units (e.g. Institute of Computer Science, Institute of Mathematics) and the Information Technology Services (ZID).

News

bootcamp

Philipp Gschwandtner is organizing a funded FFG Digital Pro Bootcamp in 2023, aimed at training participants from industry in performance optimizations of their specific applications. Further details can be found here (in German).

PRACE Logo

Long live PRACE! While new HPC access proposals for Tier-0 supercomputers are now handled by EuroHPC, the Research Center HPC was able to get several PRACE access projects awarded since 2018, with a total of 318 million core hours granted. Specific projects include Climate research in mountainous regions, antibody structure prediction, and simulation of gamma-ray emissions of binary star systems. New EuroHPC proposals are already being prepared and are partially already under review.


Nach oben scrollen