[FFmpeg-devel] Fix VP3 IDCT on Win64

[Reimar Döffinger]

On Thu, Aug 26, 2010 at 10:50:52PM +0200, Michael Niedermayer wrote:
> > - it is much less likely to be an issue since (most?) compilers
> >   do not use MMX instructions and very rarely keep values stored
> >   in the FPU (in contrast to SSE where they do both, even more so
> >   on Win64).
> 
> we dont mix float and mmx code without emms
> and before you start, we dont support compilers that create float asm out of
> integer C and noone else (kernel comes to mind) will support such compiler
> either

You misunderstood, I specifically meant compilers that optimize the whole
program at once and thus could place float ops at completely different
places in the binary from where they are in the source.
I know this is a very theoretical problem, however over the years
compilers have made a lot of issue real that were considered purely
theoretical before as well (strict aliasing rules for example).
Not saying this is something that "needs" to be fixed though.