Tuesday, March 23

State-of-the-Art FFT -- Algorithms, Implementations, and Software

10:00 AM-12:00 PM
Room: Executive Salon 3

The fast Fourier transform (FFT) is a crucial tool in many areas of scientific computing. Work on fully general algorithms, high-performance implementation, and good quality software packages with smooth interface and comprehensive functionality is a key contribution. The speakers in this minisymposium will discuss high performance and parallel implementation of FFT, mathematical encapsulations of FFT algorithms that are amenable to automatic implementations tuned to hardware platforms, and a survey of existing software, including discussion of the lack of standard functionalities.

Organizer: Ping Tak Peter Tang
Lawrence Berkeley National Laboratory

10:00-10:25 A High Performance Parallel Implementation of the 3-D FFT Algorithm
Ramesh Agarwal, IBM T. J. Watson Research Center
10:30-10:55 Automatic Generation and Implementation of FFT Algorithms
J. R. Johnson, Drexel University; and R. W. Johnson, Math Star, Inc., Minneapolis
11:00-11:25 Parallel Structured Prime Length FFTs
Charles Sidney Burrus, Rice University; and Ivan W. Selesnick, Polytechnic University
11:30-11:55 A Proposal for a Comprehensive Package for Discrete Fourier Transform
Ping Tak Peter Tang, Organizer

PP99 Home


Program Updates

Speaker Index




MMD, 1/5/99