Re: Odd behavior of time-domain convolutions

[Quentin Spencer]

Fredrik Lingvall wrote:
> Perhaps slightly OT, but anyway, as a part of a toolbox for modeling 
> acoustic waves
> (http://www.signal.uu.se/Toolbox/dream) I have a few functions for 
> computing (large)
> convolutions (the routines are threaded and can handle matrix inputs) 
> both in the time
> and the frequency domains (using fftw). The freq. domain algorithm is 
> expected to be much
> faster than the time-domain version for long convolutions. The 
> algorithms is implemented
> as C mex-files for  Matlab and C++ oct-files for Octave.  The odd 
> thing is that the  Matlab
> time-domain version is much faster than the time-domain Octave version 
> even though
> the CFLAGS and the CXXFLAGS are identical (gcc 4.1.2 was  used). See 
> the (embedded) plots below:

Is the underlying code the same, just with a mex/oct wrapper around it, 
or is it two completely separate implementations? I've used the wrapper 
approach before where the only difference was the interface code, and 
even then the mex code was a little faster, but not by the large margins 
you are observing.

Quentin