Re: libmoto PowerPC-optimized math library?

[Chris Baker <electro@1939worldsfair.com>]

Michel Dänzer <daenzerm@student.ethz.ch> writes:

> Hadess wrote:
> 
> > > > Of note is the fact that there is very little difference between
> > > > the benchmark results with no extra math library and with the
> > > > libmoto library except for the Fourier transform test, where
> > > > there's a pretty dramatic difference (improvement of about 3X with
> > > > libmoto over no special library).
> > >
> > > This rings a bell here. "MP3 encoding" :)

I always use the `high quality' setting (-h in lame).  I figure that I
spend more time listening than encoding.

> > My bell says images and video (and also mp3, sure =). Although I don't know
> > which algorithm is used in most of these formats. Fourier or Huffman.
> 
> MP3 encoding certainly uses Fourier, the question is how important it is...
> 
> I've tried building lame 3.83 with libmoto, and it seems to make no real
> difference :(

If you look at fft.c you'll notice that they use lookup tables.  This
is ironic, of course, because this approach is even *less* accurate
(and a lot faster) than whatever they're doing in libmoto.  The point,
I guess, is that if you want to trade speed for accuracy, you can do
it just fine without Motorola's help :-)

> I've just replaced -lm in the Makefiles by -lmoto, and I've noticed that the
> binary is still dynamically linked against libm - is there anything else I
> have to do for libmoto to be used?

I was wondering about this, too.  Any gcc gurus out there?

Cheers,

cbb