Newsgroups: comp.parallel.mpi
From: cesari@inf.ethz.ch (Giovanni Cesari)
Subject: parallel 1D fft using MPI
Organization: Dept. Informatik, Swiss Federal Institute of Technology
Date: 25 Jul 1996 07:49:53 GMT
Message-ID: <4t78v1$53t@neptune.ethz.ch>

Hello,

I'm writing a package for parallel multiple precision arithmetic.
To perform multiplication I use several method. One of them is based
on real fft.
The package is written in C++ using MPI.
Does anybody has a good implementation of a parallel 1D real fft 
which I could use ?

Moreover, does anybody has experience with floating point error using fft ?
At the end of the fft I have infact to reconstruct an integer array
taking only the correct bits.

Thank you very much for your help.

Giovanni


=============================================================================
Giovanni Cesari
Institute for Theoretical Computer Science      Phone: +41/1/632 6766
ETH-Z                                           fax:   +41/1/632 1172
8092 Zurich, Switzerland                        e-mail:cesari@inf.ethz.ch
=============================================================================






