Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.
|
|
|
|
Re: fftw tuning
Posted:
Mar 8, 2011 1:05 PM
|
|
In article <0df27dd3-9223-4fc5-ba5a-e46a3bca05cb@b13g2000prf.googlegroups.com>,
Felipe G. Nievinski <fgnievinski@gmail.com> wrote:
>The documentation for Matlab's fftw interface states that:
>
> fftw('planner', method) sets the method by which the tuning
>algorithm searches for a good FFT algorithm when the dimension of the
>FFT is _not_ a power of 2.
>
>So when the input dimension _is_ a power of 2 there is no potential
>speed up in tuning the FFTW library? I seem to be getting different
>performance for different tuning even with power-2 input.
No. The speed attainable depends on the dimension of the FFT,
and the algorithms differ considerably accordingly.
FFT's for sizes with many small factors (144, 180) can have
a speed advantage over comparably sized powers of 2.
<SNIP>
>
>Thanks,
>-Felipe.
--
--
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst
|
|
|
|
