TU Wien Informatics

About

Parallel Computing deals with the efficient utilization of parallel processing resources for the solution of computational problems. This sounds dry, but since all modern, general purpose computing devices are in some way or the other parallel computers, parallel compting is ubiquitous and inevitable.

Since not all computational problems are easily amenable to being solved in parallel, it is fascinating and challenging, and abounds with issues and problems that must be resolved better. Parallel Computing at TU Wien focuses on efficient utilization and modeling of real, existing architectures and systems (shared-memory multi-cores, distributed memory systems, hybrid and accelerated systems), on algorithms, interfaces, libraries and, to some extent, applications; and with idealized models of parallel computuations to explore the limits of parallelization.

The research area has specific expertise and interest in message-passing parallel computing, interfaces like MPI, benchmarking of parallel algorithms, scheduling, shared-memory algorithms and data structures, and parallel algorithms. All these topics are dealt with extensively in lectures offered by the research division.

The research Unit Parallel Computing is part of the Institute of Computer Engineering.

Sascha Hunold
Sascha Hunold S. Hunold

Associate Professor
Assoc.Prof. Dipl.-Inf. Dr.

Jesper Larsson Träff
Jesper Larsson Träff J. Träff

Head of Research Unit
Univ.Prof. Dr. / MSc PhD

2021

  • Teaching Complex Scheduling Algorithms / S. Hunold, B. Przybylski / Talk: 11th NSF/TCPP Workshop on Parallel and Distributed Computing Education (EduPar 2021) in conjunction with 35th IEEE IPDPS 2021 - Online Conference, Portland, Oregon, USA; 2021-05-17 - 2021-05-21; in: "IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPS Workshops 2021", IEEE, (2021), ISBN: 978-1-6654-1192-9; 321 - 327
  • POSTER: A more Pragmatic Implementation of the Lock-free, Ordered, Linked List / J. Träff, M. Pöter / Poster: 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2021) - Online Conference, Seoul, South Korea; 2021-02-27 - 2021-03-03; in: "Proceedings of the 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2021)", J. Lee, E. Petrank (ed.); ACM, (2021), ISBN: 978-1-4503-8294-6; 457 - 459

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

 

  • Jesper Larsson Träff: Innovation Radar: Innovation Title: PGAS-based MPI with interoperability; Innovation Category: Exploration; FP 7 project EPiGRAM
    2018 / Project
  • Sascha Hunold: Best Paper Award: "Reproducible MPI Micro-Benchmarking Isn't As Easy As You Think", S. Hunold, A. Carpen-Amarie, J. Träff, 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014, Kyoto, Japan, September 9-12, 2014
    2014 / Program Chairs of EuroMPI/ASIA 2014 / Japan
  • Jesper Larsson Träff: Best Paper Award: "Reproducible MPI Micro-Benchmarking Isn't As Easy As You Think", S. Hunold, A. Carpen-Amarie, J. Träff, 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014, Kyoto, Japan, September 9-12, 2014
    2014 / Program Chairs of EuroMPI/ASIA 2014 / Japan
  • Bartek Przybylski, Poland / Title Scheduling generalized unit-time jobs on two parallel machines Abstract In many practical cases the processing time of a job may change in reaction to environmental factors, such as the amounts of resources available, the number of jobs finished earlier (position-dependent scheduling), the amount of work already performed (sum-of-processing-times-based scheduling) or the starting time of a job (time-dependent scheduling). There are not many results on parallel-machine scheduling of jobs with variable processing times, and most of the results in this area concerns jobs without precedence constraints. We consider parallel-machine scheduling problems in which precedence-constrained jobs have variable processing times. After introducing a model of integral-based processing times, we will show how it relates to position-dependent unit-time jobs. Then, we will present a number of results on parallel-machine scheduling of generalized unit-time jobs.
  • Bartek Przybylski, Poland / Adam Mickiewicz University, Poznan
  • Martti Forsell, Finland / from VTT (VTT Technical Research Centre of Finland)
  • Nikita Koval, Austria

Soon, this page will include additional information such as reference projects, conferences, events, and other research activities.

Until then, please visit Parallel Computing’s research profile in TISS .