[Mplayer-cvslog] CVS: main/postproc swscale.c,1.21,1.22 swscale.h,1.1,1.2

[Michael Niedermayer]

Hi

On Monday 29 October 2001 19:36, Nick Kurshev wrote:
...
> > > 2.
> > > It would be more readable to have:
> > > -			MOVNTQ(%%mm0, (%4, %%eax, 4))\
> > > -			MOVNTQ(%%mm2, 8(%4, %%eax, 4))\
> > > -			MOVNTQ(%%mm1, 16(%4, %%eax, 4))\
> > > -			MOVNTQ(%%mm3, 24(%4, %%eax, 4))\
> > > +			MOVNTQ" %%mm0, (%4, %%eax, 4)\n\t"\
> > > +			MOVNTQ" %%mm2, 8(%4, %%eax, 4)\n\t"\
> > > +			MOVNTQ" %%mm1, 16(%4, %%eax, 4)\n\t"\
> > > +			MOVNTQ" %%mm3, 24(%4, %%eax, 4)\n\t"\
> > > (Althrough it's looks like cosmetic)
> >
> > i dont see why this would be better
>
> Simply because it produces additional warning - your MOVNTQ redefines
> MOVNTQ from ../mmx_defs.h But if you want - I'm ready suppress every
> warning from this stuff. (If yes - then please don't commit everything
> related with this stuff during next day) [snip]
i fixed the warning
 
>
> > not ok, because it might slow things down, and i never touch data after
> > movntq so i doubt it would have any effect
> >
> >and even if i would touch it, it would be ok, i checked intels software
> >developer manual because i wasnt 100% certain about it.
> >sfence is only required if something else than this cpu accesses the data,
> >for example an other cpu or a pci card via busmastering
>
> It's not clear from Intel's documentation.
> From other point this code will be run on K7 too ;)
yes
> Simply it should be a good habit, imho ;)
there is an sfence at the end of the swscale function, and i dont touch data 
which has been written with non temporal mov´s so how could there be a 
problem?!?!?!

Michael