Saturday, March 15

4:00 PM-6:00 PM
Greenway F-H

MS14
Fast Parallel Orthogonal Transforms: Theory, Implementation, and Applications

The use of Legendre functions and spherical basis functions are critical in many applications requiring computations on spherical domains. Some examples are weather, ocean, and climate modeling. Other applications can be found in antenna design and in physics. Traditionally, some of the problems have been "coerced" onto Cartesian grids because of the existence of fast computational techniques, such as the FFT, for such grids.

Until recently, algorithms for Legendre functions and spherical harmonic transforms were of arithmetic complexity O(NxN). Driscoll and Healy discovered an exact method of arithmetic complexity (N log₂ N). The work presented in this minisymposium covers several aspects of this basic algorithm, in particular, issues of parallelization.

Organizer: S. Lennart Johnsson
University of Houston and Harvard University

4:00 Orthogonal Polynomial Transforms

David Maslen, Utrecht University, The Netherlands

4:30 Hierarchical Load-Balancing for Parallel Fast Legendre Transforms

Nadia Shalaby, Harvard University and S. Lennart Johnsson, Organizer

5:00 Applications of Fast Orthogonal Polynomial Transforms

Dan Rockmore, Dartmouth College

5:30 Parallel FFTs and FMMs in a Global Shallow Water Model

Ruediger Jakob-Chien, University of Colorado, Denver

PP97 Homepage | Program Updates|
Registration | Hotel Information | Transportation | Program Overview | Program-at-a-Glance