While the purpose of SciDAC is scientific discovery, due to the multiscale nature of complex phenomena this is tantamount in many areas of science to harnessing new levels of performance on data sets of new scale. Computational platforms at the petascale should be coming on line in 2008, and our mathematical and software infrastructure must target this scale—millions of processor cores—in anticipation. Consequently, a major theme of SciDAC 2007 will be the identification of challenges for specific leading applications at the petascale, and approaches for overcoming these challenges. Application experts will be asked to focus on performance and scaling issues. Enabling technologies experts will be asked to focus on petascale applications and architectures.

Approximately three hundred researchers from across the SciDAC program and related endeavors will assemble in Boston. Through invitation-only talks and posters, they will offer updates that span the diverse areas of SciDAC. SciDAC 2007 will also celebrate the achievements of computational science worldwide. Computational science leaders from abroad will join in offering additional updates in important areas relevant to understanding our universe on its largest and smallest scales, from understanding Earth's climate to developing new energy sources. Astrophysics, biology, combustion, fusion, materials, nanotechnology, particle physics, and subsurface flows are among the applications that will be featured. Enabling technologies presented at the meeting will include mesh generation for complex geometry, advanced discretizations, problem adaptability, partitioning and balancing tools, optimal solvers, data archiving and data mining, software componentization, performance engineering, and compilation techniques for advanced architectures.