oasys-kit / fftw Goto Github PK

This is FFTW, a collection of fast C routines to compute the Discrete
Fourier Transform in one or more dimensions.

`OFFICIAL' CODE:

The doc/ directory contains the manual in texinfo, postscript, info,
and HTML formats.  Frequently asked questions and answers can be found
in the FAQ/ directory in a variety of formats (including HTML).

The fftw/ directory contains the source code for the complex transforms,
and the rfftw/ directory contains the source code for the real transforms.

Large portions of the source are automatically generated by a program
in the gensrc/ directory (written in Objective Caml).  You do not need
this program to use FFTW, since FFTW comes with a default set of
pregenerated codelets.  You are, however, welcome to look at and play
with the generator (see the FFTW manual for more information).

The threads/ directory contains an parallel version of FFTW (for
shared-memory machines) that uses threads.  See the "Multi-threaded
FFTW" section of the manual for more information.

The mpi/ directory contains a parallel version of FFTW for transforms
on machines with MPI.  (This code, unlike our other two parallel
transforms, supports distributed memory machines.)  See the "MPI FFTW"
section of the manual for more information.

fortran/ contains some constant definitions for using FFTW from
Fortran (see the FFTW manual), and also a small example program.

Installation instructions are provided in the manual (don't worry, it
is straightforward).

`UNOFFICIAL' CODE (for you to play with):

matlab/ contains code that allows you to call FFTW from MATLAB.

The cilk/ directory contains an parallel version of FFTW written in
Cilk.  Cilk is a cool C-like language in which you can write spawn
foo() : foo will be executed in parallel with the main thread and the
cost of spawn is just a few cycles (compare this with all the mess you
have to do to create a posix thread and pay 3000 cycles for it).  More
info on Cilk can be found at http://supertech.lcs.mit.edu/cilk/.

CONTACTS
--------

FFTW was written by Matteo Frigo and Steven G. Johnson.  You can
contact them at [email protected].  The latest version of FFTW,
benchmarks, links, and other information can be found at the FFTW home
page (http://www.fftw.org).  You can also sign up to the fftw-announce
mailing list to receive (infrequent) updates and information about new
releases; to do so, go to:

	http://www.fftw.org/mailman/listinfo/fftw-announce