[Ffmpeg-devel] [PATCH] SSE counterpart of ff_imdct_calc_3dn2

[Guillaume POIRIER]

Hi,

On 8/23/06, Luca Barbato <lu_zero at gentoo.org> wrote:
> Guillaume POIRIER wrote:
> > Hi
> >
> > Which processor did you test it with? I bet the last loop can be
> > faster on SIMD-friendly processors such as P4 or Core2/conroe
>
> what about a G5?

Ahem... well, I was specifically talking about Zuxy's patch... on x86
arch, as each CPU benefits from SIMD optimization in a different way
(some cpu prefer mmx or sse).

After re-reading the thread, I noticed that Zuxy tested on a P-M,
which AFAIK isn't as strong as P4 on SSE...

Now about the 970... I don't have enough experience to comment on
that... just that's it's way simpler to program for Altivec as the
instruction set is lighyears ahead of x86 Simd's in terms of
coherency.

Guillaume
-- 
A thing is not necessarily true because a man dies for it.
-- Oscar Wilde