MS38/Grande Ballroom
Parallel Tools and Algorithms for Large Scale Computations
Software development effort continues to be a prohibitive cost in the use of parallel computers. In this minisymposium, the speakers explore different methods for easing this burden. In particular, the speakers have been developing specialized tools for certain classes of problems, including portable packages, class libraries, and some re-design of algorithms that extends over a range of architectures. They will concentrate on applications using sophisticated numerical algorithms such as adaptive mesh refinement and domain decomposition. 
Organizer: Marsha J. Berger  
Courant Institute of Mathematical Sciences, New York University 
8:00	Portable Scientific Programming under LPARX
Scott Baden and Scott R. Kohn, University of California, San Diego
8:30	Parallel Adaptive Mesh Refinement on Shared Memory Architectures 
William Y. Crutchfield, Lawrence Livermore National Laboratory 
9:00	C++ Classes for Architecture-Independent Parallel Adaptive Mesh Refinement
Daniel J. Quinlan, Los Alamos National Laboratory
9:30	PETSc: a Portable Extensible Toolkit for Parallel Scientific Computing
Barry Smith, University of California, Los Angeles