FFTW

General Information

Short Description: Routines for discrete Fourier transforms

Long Description:
FFTW is a collection of C routines for computing the discrete Fourier transform. A parallel version using mpi also is implemented.

Category: Mathematics and Statistics

Vendor/Developer:

Vendor Website: http://www.fftw.org/

Software Website: http://www.fftw.org/doc/

Programming Language:

Terms of Use:

Installation Information

Version: 3.1.2      Version Date: 0000-00-00

Installed Site: National Center for Supercomputing Applications

Installed Platform: Intel 64 Linux Cluster

Point of Contact: Greg Bauer

Installation Status: Complete

Installation Date: 2007-12-22

Usage Information

Usage URL:

Usage Information:

FFTW is installed in /usr/apps/math/fftw on the Abe cluster.

Version 3.1.2

Version 3.1.2, which currently does not have MPI enabled FFT routines, is built with the Intel compiler and the GNU compiler.

To use FFTW 3.1:

Intel compiler: soft add +fftw-3-intel

GNU compiler: soft add +fftw-3-gcc

Compiling and Linking:
serial ifort -I ${FFTW_HOME}/include -o testf testf.f -L ${FFTW_HOME}/lib -lfftw3
threaded ifort -I ${FFTW_HOME}/include -o testf testf.f -L ${FFTW_HOME}/lib -lfftw3_threads -lfftw3

For single precision, use: -lffw3f

Please see the User Guide for more information in using FFTW in your application.

NOTE: New versions of Intel's Math Kernel Library (MKL) provides a wrapper to FFTW3 so that FFTW users can also try using MKL's DFTI routines by recompiling and linking. See the software repository entry for MKL for more information.

User Guide Homepage:

License Information

Number of Licenses:

User Fee:

Type of License: N/A

Type of Installation: Available to all allocated users

Distribution Information: