Go to the first, previous, next, last section, table of contents.


Matteo Frigo was supported in part by the Defense Advanced Research Projects Agency (DARPA) under Grants N00014-94-1-0985 and F30602-97-1-0270, and by a Digital Equipment Corporation Fellowship. Steven G. Johnson was supported in part by a DoD NDSEG Fellowship, an MIT Karl Taylor Compton Fellowship, and by the Materials Research Science and Engineering Center program of the National Science Foundation under award DMR-9400334.

Both authors were also supported in part by their respective girlfriends, by the letters "Q" and "R", and by the number 12.

We are grateful to SUN Microsystems Inc. for its donation of a cluster of 9 8-processor Ultra HPC 5000 SMPs (24 Gflops peak). These machines served as the primary platform for the development of earlier versions of FFTW.

We thank Intel Corporation for donating a four-processor Pentium Pro machine. We thank the Linux community for giving us a decent OS to run on that machine.

The genfft program was written using Objective Caml, a dialect of ML. Objective Caml is a small and elegant language developed by Xavier Leroy. The implementation is available from ftp.inria.fr in the directory lang/caml-light. We used versions 1.07 and 2.00 of the software. In previous releases of FFTW, genfft was written in Caml Light, by the same authors. An even earlier implementation of genfft was written in Scheme, but Caml is definitely better for this kind of application.

FFTW uses many tools from the GNU project, including automake, texinfo, and libtool.

Prof. Charles E. Leiserson of MIT provided continuous support and encouragement. This program would not exist without him. Charles also proposed the name "codelets" for the basic FFT blocks.

Prof. John D. Joannopoulos of MIT demonstrated continuing tolerance of Steven's "extra-curricular" computer-science activities. Steven's chances at a physics degree would not exist without him.

Andrew Sterian contributed the Windows timing code.

Didier Miras reported a bug in the test procedure used in FFTW 1.2. We now use a completely different test algorithm by Funda Ergun that does not require a separate FFT program to compare against.

Wolfgang Reimer contributed the Pentium cycle counter and a few fixes that help portability.

Ming-Chang Liu uncovered a well-hidden bug in the complex transforms of FFTW 2.0 and supplied a patch to correct it.

The FFTW FAQ was written in bfnn (Bizarre Format With No Name) and formatted using the tools developed by Ian Jackson for the Linux FAQ.

We are especially thankful to all of our users for their continuing support, feedback, and interest during our development of FFTW.

Go to the first, previous, next, last section, table of contents.