Parallel Computing 33 (2007) 81–82 www.elsevier.com/locate/parco
Guest editorial
Trends in parallel computing
Access to high-performance computers has increased throughout the world as a result of advancements in hardware and software. The worldwide focus on computational science, therefore, has brought people together from different continents in the past decade. An example of this is a series of International conferences on Computational Science and Engineering (ICCSE) (2004 – China, 2005 – Turkey, and 2006 – USA) we have organized. The first conference was hosted by Nankai University – China and the third one is scheduled to be hosted by the State University of New York at Brockport. The second of ICCSE conferences was held in Istanbul, Turkey on June 27–30, 2005. The mission of the ICCSE series is to gather world’s best minds in parallel computing and computational science to explore the frontiers of applying supercomputing to life and physical sciences and engineering. The conference included invited talks in the morning and multiple sessions in the afternoon on referred research papers. The conference was hosted on the campus of Istanbul Technical University (ITU) and jointly sponsored by ITU, the State University of New York (SUNY), with support from the Scientific and Technological Research Council of Turkey (www.tubitak.gov.tr), US National Science Foundation, Turkish Government State Planning Organization (www.dpt.gov.tr), IBM, Intel, HP, and Isß Bankasi. Although the theme of the conference was computational science and applications, the emphasis was molecular dynamics and computational chemistry, the areas where most of the invited speakers came from. The contributions of the invited speakers on the latest developments in their respective field were instrumental to the success of the conference. We believe that the conference made important contribution to the computational science and engineering field by bringing from around the world together the scientists, application people, tool developers and hardware developers. There were 14 different countries from which the accepted papers were submitted and presented. In the first day, the opening keynote addresses were given by Horst Simon, Jack Dongarra, and Tim Clark. Horst Simon described the Top Three Breakthroughs of past 20 years as well as Top Three Challenges of the next 20 years. Jack Dongarra talked about a new generation of software libraries and algorithms for distributed and parallel environments. Alan Gara (IBM) and Yuefan Deng (SUNY-Stony Brook) talked about IBM BlueGene/L machine. On the education side, Rubin Landau described his undergraduate program in computational physics while Osman Yasßar presented computational science both as a strategy for higher education to improve technical workforce and as pedagogy for secondary schools to teach mathematics and science. In addition to the formal technical program, there was ample opportunity for informal discussions and working group meetings during the generous break periods. Papers submitted to the conference were reviewed in two steps, those deemed worthy for conference presentation and a selected few for consideration in a special issue of this journal. There were over 65 papers submitted with a 45% rejection rate. Each paper was reviewed by at least three referees. We sincerely thank to our scientific committee members for such a superb job. Out of 42 accepted papers, nine were selected for consideration in a special issue of Parallel Computing. The selection was based solely on the referees’ suggestions. Reviewers who attended the conference were asked to provide an evaluation of pre-selected papers for the special issue. Authors were asked for a revised version of their papers for a final review by guest editors and their team. This review process took more than 6 months after the conference. Following an additional final review 0167-8191/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.parco.2006.11.001
82
Guest editorial / Parallel Computing 33 (2007) 81–82
by the journal editors, we are presenting 5 papers as representative of the breadth of the technical content of the Conference. In this issue, the reader will see several applications of parallel computing, including an option pricing method to predict financial performance, molecular dynamics computations on a supercomputer built for quantum chromodynamics, a 3-D finite difference method for electro-magnetodynamics, and a fault diagnosis system for airline engines. The reader will also witness development of mathematical libraries and software tools for more efficient storage of sparse matrices and preconditioners for distributed computational grids. Majority of the techniques and tools described for presented applications are believed to be applicable to other areas. We hope that the readers will engage in discussions with authors to advance research techniques and disseminate the results. We hope that the readers find these papers useful and informative and will consider participating in future ICCSE workshops. Osman Yasßar Department of Computational Science, State University of New York College at Brockport, Brockport, NY 14420, USA E-mail address:
[email protected] Hasan Dag˘ Computational Science and Engineering Program, Istanbul Technical University, Maslak, Istanbul, Turkey E-mail address:
[email protected] Available online 12 January 2007
