The researchers have received over 36 million to finance the project. Even today’s supercomputers made up of tens to hundreds of computing cores that make up a computer system with massive parallelism. In order to effectively utilize this capacity, access to efficient and scalable parallel algorithms and software, which has long been an area within the computation-oriented and computer science research at Umeå University.
HPC2N (High Performance Computing Center North) at Umeå University offers, as part of VR’s metacentre SNIC, national access to first-class supercomputer system for large-scale computing and mass storage of data.
A trillion operations per second
Future supercomputers will be even more extreme parallel, scalable and heterogeneous; the goal is to deliver supercomputer with a computing capacity of a trillion operations per second (1 trillion = 1 000 000 000 000 000 000; 10 raised to 18) by 2020.
This dramatic developments in turn places new and challenging demands for efficient numerical algorithms and software libraries. The purpose of NLAFET is to tackle these challenges and ultimately deliver new scalable numerical library for large-scale fundamental problems in numerical linear algebra, including the solution of dense and sparse systems of equations and eigenvalue problems.
The components of a program library can be seen as Lego bricks used to build software for different application areas. Today these applications in virtually all of the faculty and research, with the longest tradition in technology and science.
Analysis of large amounts of data
> In addition to simulations of physical models of reality, processing and analysis of large amounts of data an important area with many new applications.
NLAFET stands for “Parallel Numerical Linear Algebra for Future Extreme Scale Systems” . It is a high-profile research projects (RIA) for extremely scalable computer systems, funded by the European Commission in the section Future and Emerging Technologies (FET) as part of Horizon in 2020.
Horizon 2020 is the EU ‘ Framework Programme for funding research and innovation, with funding of almost € 71 billion over seven years (2014-2020).
In order to achieve the project objectives requires a broad focus on developing new algorithms, exploring advanced scheduling strategies and execution systems, automatic optimization of software (so-called offline and online auto-tuning), and not least to avoid bottlenecks related to communication and synchronization of the millions of computational tasks that together form the whole.
Preconditions for creativity
NLAFET project coordinated by Umeå University, with international partners from the University of Manchester (UniMan), United Kingdom, Institut National de Recherche one Informatique et Automatique (INRIA), France and Science and Technology Facilities Council (STFC), United Kingdom.
Recently held the project kick-off conference at Umeå University with participants from all four parties. The conference was organized jointly by the Department of Computer Science and HPC2N. At the conference, the arrangement of the research project and its various components.
A key challenge for NLAFET and for all projects within the Horizon 2020 is to create the best possible conditions for creative research to effectively collaborate and together achieve the high ambitions of the project.
CONTACT
Bo Kågström, professor, chief coordinator and scientific director of NLAFET Department of Computer Science and HPC2N, Umeå University. Phone: 073-620 54 19. E-mail: bokg@cs.umu.se
No comments:
Post a Comment