This NSF award will provide partial financial support for the Twentieth International Conference on Domain Decomposition Methods, to be held at the San Diego Supercomputer Center on the campus of the University of California at San Diego, in February 2011. Domain decomposition, in some form, is by far the most common paradigm for large-scale simulation on massively parallel distributed, hierarchical memory computers. This paradigm permeates almost all arenas of simulation, from such practical concerns as numerical climate modeling, modeling of complex biological molecules and biological processes, the design and testing of automobiles, aircraft and other structures, and the modeling of complex porous media flows, to more abstract scientific endeavors such as modeling black holes and searching for the origins of the universe. However, many users of domain decomposition do not avail themselves of the most sophisticated and efficient methods known. This has become more pronounced as the size of the large-scale scientific simulation community has grown with the availability of hardware in the past few years, particularly inside scientific communities without traditionally strong ties to the computational mathematics or computer science communities.
The goal of this conference is to promote advances in domain decomposition methods by encouraging interdisciplinary technical interchange throughout the international computational science and engineering communities, and by taking advantage of conference siting, to attract to this important area many new U.S. scientists. It is strongly aligned with NSF's, and more generally, our national priorities to promote and advance fundamental interdisciplinary research in Computational Science and Engineering. NSF funding will make possible financial support for graduate students, post-docs, and other young researchers who otherwise might not be able to attend. The funding will also allow for increasing participation of women, African Americans, and Hispanic Americans from across the nation as well as from the diverse Southern California scientific community.
(DD20) was held February 7--11, 2011 at the San Diego Supercomputer Center on the Campus of the University of California, San Diego. Domain decomposition, in some form, is by far the most common paradigm for large-scale simulation on massively parallel distributed, hierarchical memory computers. Conferences in this series have been organized in twelve countries since the first meeting in 1987. With DD20, the series returned to the USA for the first time since 2005, when it was held at Courant Institute in New York City. DD20 attracted 199 registered participants drawn form the fields of numerical analysis, applied mathematics, computer science, and application areas that drive acquisitions of the largest parallel computers, especially in areas modeled by partial differential equations, such as fluid dynamics, structural mechanics, biomechanics, geophysics, plasma physics, astrophysics, radiation transport, electricity & magnetism, porous media, and the like. The conference featured 176 technical presentations (invited, mini-symposium and contributed lectures, and poster presentations). The book of abstracts is a PDF file with hyperlinks; it was placed on line long before the the opening of the conference, and remains available for download on the DD20 webpage. http://ccom.ucsd.edu/~dd20/ On February 6, a one-day introductory short course on Domain Decomposition was given by Professor Olof Widlund from Courant Institute of New York University. The short course attracted about 40 participants. The slides and other course materials remain available for download on the DD20 webpage. The conference proceedings volume will be published by Springer in the series ``Lecture Noted in Computational Science and Engineering.'' We have received 86 submissions for this volume, that are in the reviewing process as this is written. As was the case for previous conferences in this series, the proceedings will also be available online at the the DD20 webpage and possibly other locations. We advertised financial support for young computational scientists interested in attending the conference, and were able to offer some support to all 25 graduate students and postdocs who applied on the DD20 webpage. In addition to NSF funding, the conference was supported by a grants from DOE, LLNL and Sandia National Laboratories, and several groups on the UCSD campus (CalIt2, SDSC and NBCR). The balance of the the funding came from registration fees.