[x264-devel] stride problem on powerpc

[Jason Garrett-Glaser]

On Wed, Sep 22, 2010 at 6:47 AM, Manuel Rommel <maaanuuu at gmx.net> wrote:
> Hello,
>
> in x264_weight_cost in slicetype.c, line 118, satd_8x8 may be called with a stride of 8. This leads to wrong calculations on powerpc because VEC_DIFF_H is tuned for an input stride of c*16 for pix1.

If that's the only place that we do that, feel free to change the C instead.

Dark Shikari