[x264-devel] commit: add AltiVec implementation of x264_pixel_var_16x16 and x264_pixel_var_8x8 (Guillaume Poirier )

[Loren Merritt]

> sqr_v = vec_add( sqr_v, vec_sld( sqr_v, sqr_v, 8 ) );
> sqr_v = vec_add( sqr_v, vec_sld( sqr_v, sqr_v, 4 ) );

vec_sums. I said that the biggest sum (16x16 squared) fits in 24 bits.

--Loren Merritt