Exascale: challenges and benefits.
Theme: H2020: ICT for Excellent science
Hall 1.1, 07/11/2013 (09.00-10.30)
This session will focus on the strategic nature of the high-performance computing and will aim to demonstrate that although the path from the present petaflop to the future exaflop era involves numerous complex challenges, the expected benefits far outweigh the costs and the efforts.
Fans and opponents of the supercomputers of all ages and occupations are welcome to attend the session.
Chair: *Marie-Christine Sawley* (Intel Corporation, France)
Links and Documents
- Jean-Yves BERTHOU (ANR, ICT, France), Developing the next generation of HPC towards exascale
The need for a holistic approach to tackle the challenges of massive parallelism, encompassing novel architectures, system software, file systems, compilers, programming environments and tools, exascale system prototyping etc.
- Catherine RIVIERE (GENCI, France), Providing access to the best facilities and services for both industry and academia
A world-class European HPC infrastructure is needed to increase competitiveness in science and industry (including SMEs).
- Michael RESCH (University of Stuttgart, Germany), Achieving excellence in HPC applications
A whole range of applications (e.g strategic scientific applications addressing societal challenges, emerging domains dealing with big data, or the emerging area of global systems science that addresses new forms of interaction between science and policy.) require optimization to fully exploit the performance of current petaflow and future exaflop systems. New computational methods and algorithms must be developed, and applications must be re-programmed in radically new ways.
- Thomas LIPPERT (Forschungszentrum Jülich, Jülich Supercomputing Centre, Germany), HPC for HBP - The Human Brain Flagship Project's Supercomputer Challenges
The goal of the human brain flagship project is to re-engineer the human brain's structure, information processing capabilities and higher-order functions on a supercomputer. Besides the implementation of Europe-wide federation for the integration of all kinds of clinical and experimental brain data, a supercomputer with a memory capacity of more than 100 PetaBytes and a compute capability of more than one Exaflop/s is required. The HBP plans a stepwise realization of such true Big Data capacities and high-end parallel capabilities taking benefit from research driven by the European HPC strategy through ETP4HPC and PRACE, the partnership for advanced computing in Europe.
- Maryline LENGERT (European Space Agency, Italy), Supporting the European industry in the development of technologies and tools for the next generation of clouds
The focus is on the cloud infrastructures and their interaction with the underlying networks, on platforms and tools to develop and deploy applications in the cloud, on cloud-based services, and on their adoption by businesses, consumers and especially by the public sector.
Session organiser: Sonia SPASOVA (European Commission, DG CONNECT, Belgium)
You must log in to add a comment.