August 27, 2014 – August 28, 2014 all-day
University of Porto. FEUP. Main Building. Room B002.
Rua Doutor Roberto Frias 378
SEE DETAILED PROGRAM OF PRESENTATIONS HERE
August 27th Action Workshop
9:00 – 9:30 Welcome
9:30 – 10:15 Invited speaker: Andrea Feltrin. Project Officer. DG Connect. European Commission.
10:15 – 11:00 Invited speaker: Stefano Markidis. Assistant Professor, KTH Royal Institute of Technology, Stockholm, Sweden.
11:00 – 11:30 Coffee break
11:30 – 13:00 Presentations
13:00 – 14:00 Lunch
14:00 – 17:00 Presentations.
17:15 – 19:30 Management committee meeting
20:30 Social dinner
August 28th Action Workshop
9:00 – 11:00 Presentations
11:00-11:30 Coffee break & Posters Session
13:00 – 14:00 Lunch
14:00 – 16:30 Presentations.
16:30 – 17:00 Summary and Workshop closing
Extended versions of papers presented in NESUS Workshop 2014 can be send for publication, after further revisions in the journal, in a special issue of the journal Simulation Modelling Practice and Theory (SIMPAT)
that is indexed in JCR.
|Andrea Feltrin (PhD), Project Officer.DG CONNECT, European Commission.Summary of the presentation: The presentation will give an overview of the ‘Future and Emerging Technologies’ (FET) programme in the Horizon 2020 EU framework programme for Research and Innovation. It will also offer details about the implementation of the FET Proactive Initiative TOWARDS EXASCALE HIGH PERFORMANCE COMPUTING in the Work Programme 2014-2015 and how this initiative fits into the European HPC strategy.
|Stefano Markidis (PhD). Assistant Professor, KTH Royal Institute of Technology, Stockholm, SwedenScientific Applications at Exascale: Challenges and RoadmapSummary of the presentation: Exascale supercomputers and the need of using them efficiently pose new challenges to scientific application developers. An unprecedented amount of parallelism will be available: current studies of exascale machines estimate a total of billion processes available for computation. On exascale machines, the synchronization of such large number of processes will constitute the main bottleneck. In fact, local delays on a single process might propagate to other processes and be amplified making the use of exascale machine ineffective. After an introduction about the exascale era for applications, we present the challenges we have identified by studying two applications from the EC-funded EPiGRAM project. We discuss the main challenges for application developers: the software and hardware unknowns, the delay in adopting in applications the new features of programming systems, the “MPI + X” dilemma and the suitability of current algorithms to exascale. A roadmap to bring scientific applications to exascale will be finally discussed.
Local Organization Contact:
Directions to Meeting place from airport.
Hotels close to FEUP: