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

[Guillaume POIRIER]

Hi

On 8/23/06, Zuxy Meng <zuxy.meng at gmail.com> wrote:
> 2006/8/21, Loren Merritt <lorenm at u.washington.edu>:
> > If you can't make an sse version that's faster than C, have you tried mmx?
> > Just take the one from 3dn2 and change pswapd to pshufw.
>
> Changing the last loop to SSE or MMX dosen't bring about significant
> speedup. Anyway I have this new patch for your review:-)

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

BTW, if someone knows if there's any organization or person who can
offer ssh access to a core2/conroe machine?

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