NLAFET – a high-profile computing research project funded with nearly €4 million

[2015-11-19] Umeå University together with three international partners have been funded with nearly 4 million euros (3,907,375) for a front-line research project focusing on novel methods and software for the future supercomputer systems. The project named NLAFET is one of the high-profile extreme-scale computing research projects funded by the European Commission within the Future and Emerging Technologies (FET) program under Horizon 2020.

Photo: NLAFET Management at Umeå University. From left to right: Sen. Lecture Lennart Edblom (Adminastrive Manager), Prof. Bo Kågström (Coordinator and Scientific Director), Asst. professor Lars Karlsson (Deputy Research Group Leader).

Today's most powerful supercomputers are composed of hundreds of thousands computing cores (CPUs and accelerators) connected in high speed networks that make up a massively parallel high performance computing (HPC) system. To effectively utilize this capacity, access to efficient and scalable parallel algorithms and software is necessary.

The future supercomputers will be even more extremely parallel; the goal is to deliver HPC systems with a capacity of 1,000,000,000,000,000,000 (10 raised to the power 18) operations per second (one exaflop) by the year 2020. Such an exa-scale HPC system will also be heterogeneous and consist of millions of compute cores. This dramatic development in turn places new and challenging demands on effective scalable numerical algorithms and software libraries. 

The purpose of NLAFET is to tackle these challenges with the goal to minimizing the gap between the peak capabilities of the hardware and the performance realized by HPC applications. Achieving this requires a co-design effort including developing novel parallel algorithms, exploration of advanced scheduling strategies and runtime systems, offline and online autotuning, as well as avoiding communication and synchronization bottlenecks.

NLAFET objectives

With a focus on extreme-scale HPC systems, the NLAFET project will deliver a new generation of computational methods, parallel algorithms and software libraries for fundamental problems in numerical linear algebra, including the solution of dense and sparse systems of equations and eigenvalue problems. Linear algebra is both fundamental and ubiquitous in computational science and its vast application areas. Many of the methodologies, functionalities and solutions developed within NLAFET will have applicability to computational science in general.

The components of a numerical software library can be seen as Lego bricks, where each brick corresponds to solving a fundamental numerical problem. The Lego bricks are used as building blocks to develop software for different application areas. Today such computational science applications appear in virtually all faculties and research areas, with the longest tradition in natural science and engineering. In addition to numerical simulations of physical models of reality, the processing and analysis of very large amounts of data (BIG DATA analytics) is an important area with many new applications.

With the desire to be able to increase the scale of the science, e.g., work on extreme-scale problem sizes and higher resolutions, and add new dimensions like second and third order effects and coupled models, there will be a continuous demand of new and more powerful HPC resources. In turn, this evolution leads to new and challenging demands for efficient numerical algorithms and parallel software libraries.

The validation and dissemination of results from NLAFET will be done by integrating new software solutions into challenging scientific applications in materials science, power systems, study of energy solutions, and data analysis in astrophysics. The deliverables also include a sustainable set of methods and tools for cross-cutting issues such as scheduling, auto-tuning, and algorithm-based fault tolerance packaged into open-source library modules.

NLAFET funding

NLAFET is an acronym for the project title “Parallel Numerical Linear Algebra for Extreme Scale Systems” and is funded by the European Commission as a RIA-project within the recent FET-HPC call under Horizon 2020. Projects in Future and Emerging Technologies are expected to initiate radically new lines of technology through unexplored collaborations between advanced multidisciplinary science and cutting-edge engineering. (RIA – Research and Innovation Action in EC nomenclature.)

NLAFET partners

NLAFET is coordinated by Umeå University (UMU), Sweden, with international partners from University of Manchester (UNIMAN), United Kingdom, Institut National de Recherche en Informatique et Automatique (INRIA), France, and The Science and Technology Facilities Council (STFC), United Kingdom.

NLAFET kick-off, November 9-10

Recently, the NLAFET kick-off conference was held at Umeå University with participants from all four partners. The conference was organized jointly by the Department of Computing Science and the High Performance Computing Center North (HPC2N). At the conference, the research project and its various components were presented and discussed. A key challenge for NLAFET, and for all projects within the Horizon 2020, is to create the best possible conditions for a creative research environment that facilitate effective collaboration to achieve the high ambitions of the project.


The attached photos are from the kick-off conference at Umeå University.

Scientific principal investigators (PIs) of the NLAFET partners. 
From lef to right: Professor Iain Duff (STFC), Professor Bo Kågström (UmU, NLAFET Coordinator), Dr Laura Grigori (director of research (DR2) at INRIA), Professor Jack Dongarra (UNIMAN).

Researchers, representing all partners, participating at NLAFET kick-off.

For more information, please contact:
Professor Bo Kågström, Coordinator and Scientific Director of NLAFET
Dept. of Computing Science and HPC2N, Umeå University, Sweden 
Phone: +46-90-786 5419

Editor: Mikael Hansson

Link to news:

Page Editor: Lena Kallin Westin