Numexas: A European project to develop the future algorithms towards the Exascale era

Numexas: A European project to develop the future algorithms towards the Exascale era

Barcelona, 17th October 2013

 

 

Numerical methods and in particular the field devoted to the simulation of multi-physics phenomena using finite element, finite volume, and discrete element methods, can be understood as the art of reducing extremely complex engineering challenges into simpler problems, while maintaining the consistency of the solution.

 

This concept remains unchanged since the days when it was formulated by pioneers in computational mechanics such as Prof. Zienkiewicz, a former UNESCO Chair of Numerical Methods in Engineering.

 

Computational engineering starts off from wicked problems such as the structural design of the largest bridges in the world, a risk to the structural integrity of a harbour’s behaviour when pounded by violent waves or extreme floods, the electromagnetic response of a commercial planes to broadband emissions, or the blood circulation in a human body affected by rare diseases. The journey then moves from formulating the physical problem towards a computer implementation capturing the essence of the problem dynamics.

 

During this journey, our task consists in working with the most powerful machines available to us. That is the reason why experts in numerical methods have such extreme requirements for computational resources, and see themselves obliged to take full advantage of current developments in computer hardware and cloud computing..

 

A group of European experts in computing engineering celebrated the official launch of the Numexas project on the 17th of October. Numexas  is tasked to tackle some of the grand challenges in engineering.

 

Numexas, “Numerical Methods and Tools for Key Exascale Computing Challenges In Engineering and Applied Sciences”, is a STREP collaborative project within the FP7-ICT programme of the European Union. The goal of Numexas  is to develop, implement and validate the next generation of numerical methods running on exascale computing architectures. We will achieve this goal by implementing a new paradigm for the development of advanced numerical methods that is able to fully exploit the intrinsic capabilities of the future exascale computing infrastructures.

 

The main outcome of Numexas are a new set of numerical methods and codes that will allow industry, government and academia to solve exascale-class problems in engineering and applied sciences on the next generation of exaflop computers, with the efficiency and ease of use as today''s state-of-the-art codes.

 

The Numexas consortium includes renowned institutions specialised in the development of numerical methods to solve scientific and engineering problems: CIMNE, the International Center for Numerical Methods in Engineering and coordinator of the project, the group IKM of the LUH, the Gottfried Wilhelm Leibniz Universitaet Hannover in Germany and the National Technical University of Athens in Greece) and institutions hosting HPC facilities and supercomputing infrastructures (the Consorci Centre de Serveis Cientifics i Academics de Catalunya, CESCA in Spain and the group HKNR in LUH). The partnership is completed with QUANTECH, an SME specialised in the development and marketing of simulation software for industrial forming processes.

 

The budget of the project of three years of duration is about 2.2 million euro, having received a funding by the EU of 1.7 million euro.

 

Webpage of the project: http://www.numexas.eu/

 

For further information:

International Center for Numerical Methods in Engineering (CIMNE)
Dr. Pooyan Dadvand (numexas@cimne.upc.edu)





Kick-off meeting group photo




Wednesday, December 13, 2017
  [ login ]

The research leading to these results has received funding from the European Community's Seventh Framework Programme under grant agreement n° 611636