<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <html> <head><title> FFTW FAQ - Section 4 </title> <link rev="made" href="mailto:fftw@fftw.org"> <link rel="Contents" href="index.html"> <link rel="Start" href="index.html"> <link rel="Next" href="section5.html"><link rel="Previous" href="section3.html"><link rel="Bookmark" title="FFTW FAQ" href="index.html"> </head><body text="#000000" bgcolor="#FFFFFF"><h1> FFTW FAQ - Section 4 <br> Internals of FFTW </h1> <ul> <li><a href="#howworks" rel=subdocument>Q4.1. How does FFTW work?</a> <li><a href="#whyfast" rel=subdocument>Q4.2. Why is FFTW so fast?</a> </ul><hr> <h2><A name="howworks"> Question 4.1. How does FFTW work? </A></h2> The innovation (if it can be so called) in FFTW consists in having a variety of composable <i>solvers</i>, representing different FFT algorithms and implementation strategies, whose combination into a particular <i>plan</i> for a given size can be determined at runtime according to the characteristics of your machine/compiler. This peculiar software architecture allows FFTW to adapt itself to almost any machine. <p> For more details (albeit somewhat outdated), see the paper "FFTW: An Adaptive Software Architecture for the FFT", by M. Frigo and S. G. Johnson, <i>Proc. ICASSP</i> 3, 1381 (1998), also available at <A href="http://www.fftw.org">the FFTW web page</A>. <h2><A name="whyfast"> Question 4.2. Why is FFTW so fast? </A></h2> This is a complex question, and there is no simple answer. In fact, the authors do not fully know the answer, either. In addition to many small performance hacks throughout FFTW, there are three general reasons for FFTW's speed. <ul> <li> FFTW uses a variety of FFT algorithms and implementation styles that can be arbitrarily composed to adapt itself to a machine. See <A href="#howworks">Q4.1 `How does FFTW work?'</A>. <li> FFTW uses a code generator to produce highly-optimized routines for computing small transforms. <li> FFTW uses explicit divide-and-conquer to take advantage of the memory hierarchy. </ul> For more details (albeit somewhat outdated), see the paper "FFTW: An Adaptive Software Architecture for the FFT", by M. Frigo and S. G. Johnson, <i>Proc. ICASSP</i> 3, 1381 (1998), available along with other references at <A href="http://www.fftw.org">the FFTW web page</A>. <hr> Next: <a href="section5.html" rel=precedes>Known bugs</a>.<br> Back: <a href="section3.html" rev=precedes>Using FFTW</a>.<br> <a href="index.html" rev=subdocument>Return to contents</a>.<p> <address> <A href="http://www.fftw.org">Matteo Frigo and Steven G. Johnson</A> / <A href="mailto:fftw@fftw.org">fftw@fftw.org</A> - 25 November 2012 </address><br> Extracted from FFTW Frequently Asked Questions with Answers, Copyright © 2012 Matteo Frigo and Massachusetts Institute of Technology. </body></html>
