slideshow 1

You are here

Universität Wien (UNIVIE)

Organization

Founded in 1365, the University of Vienna (Universität Wien) is among the oldest universities in Europe. The University of Vienna offers more than 135 Bachelor, Master, Diploma, and Doctoral programmes. Research and academic programs that are offered at the University of Vienna include: computer science, economics, humanities, jurisprudence, social and natural sciences, and theology. At present about 85,000 students from more than 130 countries are enrolled.

Expertise

The Research Group Scientific Computing (SC) at the University of Vienna has achieved inter-national recognition in the field of software for parallel and distributed systems. Research focus-es on the design and development of languages, software tools, and programming environments with the goal of making distributed and parallel computing platforms accessible to a broad range of users in science and industry. SC has contributed in the past to more than ten EU projects related to high performance computing and Grid computing. Among other projects, SC has coordinated the Esprit Project HPF+ (Optimizing HPF for Advanced Applications) where they developed a source-to-source parallelization system that translates HPF+ programs into parallel F90/MPI message-passing programs. SC is currently the coordinator of the FP7 STREP project PEPPHER (Performance Portability and Programmability for Heterogeneous Many-core Architectures), which aims at developing a unified framework for programming and optimizing applications for architecturally diverse, single-node, heterogeneous many-core processors to ensure performance portability.

Role

UNIVIE will lead the overall design of the performance tuning framework (WP2). It will provide its extensive experience on HPC language, compiler and tool development to the project. UNI-VIE will also contribute to the development of tuning support for hybrid HPC architectures based on GPU accelerators. In this context, UNIVIE will bring to the project its experience and know-how gained in the EU project PEPPHER which focuses on programmability and performance portability of single-node, heterogeneous many-core systems.

Key personnel

Prof. Siegfried Benkner is the head of the research group of Scientific Computing at the University of Vienna, Austria, which currently comprises 25 people. He received MSc and Ph.D. degrees in Computer Science from the Vienna University of Technology. Siegfried Benkner contributed to several EU projects, including PPPE, PREPARE, HPF+, GEMSS, @neurIST, and was the Technical Director of the LTR project "HPF+". His research interests include parallel and distributed computing, Service-oriented software architectures, Grid and Cloud computing as well as languages, compilers and runtime systems for parallel and distributed systems. A current research focuses is on programming support for heterogeneous many-core systems in the context of the EU PEPPHER project, which is coordinated by Benkner‘s group. Siegfried Benkner has published some 100 peer-reviewed publications and is a member of the ACM and the IEEE.

Result dissemination and exploitation

UNIVIE has a strong interest in auto-tuning technologies, which will be indispensible not only for future HPC architectures, but for heterogeneous multi-core systems across the whole range of the IT spectrum in general. UNIVIE will focus on the exploitation of performance analysis and auto-tuning mechanism for GPU-based HPC systems. A goal is to adopt major developments of the AutoTune project for UNIVIE‘ s tool-chain that targets heterogeneous many-core architectures. As an academic partner, UNIVIE will put a strong focus on disseminating the results of the AutoTune project at international conferences and in scientific journals. UNIVIE will foster synergies between the EU project PEPPHER and the AutoTune project. Moreover, UNIVIE will integrate topics covered by the AutoTune project into its teaching activities at the Master‘s and Ph.D. levels.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer