Re: new diff, also improving sse16_c

> -	    s += sq[pix[7]];
> +	    register unsigned long x=*(unsigned long*)pix;
> +	    s += sq[x&0xff];

You have somewhat forget that not all architectures are little endian.
That's why these are written generic.

For various platforms there are specifily written routines.
But for C-code it will have to work everywhere

--

-- 
  .''`.    Zdenek Kabelac  kabi <at> {debian.org, users.sf.net, fi.muni.cz}
 : :' :          Debian GNU/Linux maintainer - www.debian.{org,cz}
 `. `'                          Overclocker's house :)
   `-              http://www.bbspot.com/News/2002/08/house.html        

-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com