November 11th-12th, 2010
Supercomputing Center of Galicia, CESGA
Santiago de Compostela

Vision and Goals

This workshop will provide a forum for discussing the future needs of computational techniques in mathematics and the contribution of mathematical techniques to the petascale/exascale computation and to new computational architectures. This workshop will deal not only with system architecture, software and tools, but also with models, algorithms and application software that will enable scientific discoveries in the future petascale/exascale computing.

The objective of this workshop is to engage researchers in applied mathematics and algorithms, in discussions with computational scientists, applications developers, and experts in system architecture, system software and tools, and programming frameworks, with the goal of:

• exploring research opportunities in Mathematics and Computational Science that the petascale/exascale supercomputers and also the new computational architectures could provide;
• exploring the characteristics of systems needed by future applications in the period 2011-2018;
• determining the position of researchers in i-MATH Project with respect to the Computational Science in this period;
• providing i-MATH Project with an identification of the most important open questions in this period: "co-design" of system architecture, system software and tools, as well as new programming frameworks, mathematical models and algorithms.

Summing up, the objective of the workshop is to involve the relevant communities in defining the elements of the R&D of Mathematics for petascale/exascale computing.

Intended for

• Any researcher in Mathematics who can benefit from high performance computing or develops new algorithms.
• Any researcher in Computational Science in order to present next-generation supercomputers which will improve the research abilities, permit to develop new areas of research, or provide a gateway to new research results.

Workshop structure

The Workshop will be organized using three main themes within a 2 day period, with interactions among the participants in each theme. The scope of the themes will cover the following elements:

• Applications needs and plans to develop petascale/exascale applications software.
• Mathematical models and algorithms.
• Development of algorithms for new computational architectures.
• Programming models and environments.
• System software and tools.
• System architecture.
• Co-design strategies.

The details of the workshop agenda are being developed but the following description represents the preliminary planning for the three major themes:

• Theme 1: Mathematical applications at Petascale/Exascale and new computational architectures.
• Theme 2: Programming needs for Mathematics.
• Theme 3: Fundamental algorithms at Petascale/Exascale.

Plenary discussion sessions at the end of the workshop will provide opportunities for the participants in all three themes to present and agree findings.

In the last half day, there will be a plenary session to summarize the findings and recommendations, focusing on what kind of R&D is needed in mathematical models, applied mathematics and computer science to cover the needs for applications on exascale architectures.

The position papers will be gathered in an i-MATH publication as a white book in the petascale/exascale computing mathematical research agenda.