theora dev - Jun 2006 - theora alpha 6 and MMX optimizations

Hi,
I have tried the latest SVN version (as of now) of ffmpeg2theora and libtheora.
Seems to me that the MMX optimizations are not enabled: my binary is about 50%
slower than the original binary ffmpeg2theora-0.16 found at
http://www.v2v.cc/~j/ffmpeg2theora/ and my binary built with the old theora-mmx,
and is slower as the pre-alpha6 theora.
However the optimizations are enabled in the configure:
------------------------------------------------------------------------
  libtheora 1.0alpha7-svn:  Automatic configuration OK.

  General configuration:

    Encoding support: ........... yes
    Floating point support: ..... yes
    Assembly optimization: ...... 32 bit x86

Am I missing something or is it correct that current MMX enabled theora is much
slower than the old theora-mmx?

Fabio

On Thu, 2006-06-01 at 15:11 +0200, Fabio wrote:
> Am I missing something or is it correct that current MMX enabled theora is
much slower than the old theora-mmx?
looks like USE_ASM was never defined in configure, 
this should be fixed in svn.

j

> looks like USE_ASM was never defined in configure, 
> this should be fixed in svn.
Thanks for the quick fix. Maybe we should release theora alpha 6.1 with this
fix...

Fabio