template<bool align> SIMD_INLINE v128_u8 BlurRow(const Buffer & buffer, size_t offset)
{
v128_u16 lo = DivideBy16(BinomialSum(Load<align>(buffer.src0 + offset), Load<align>(buffer.src1 + offset), Load<align>(buffer.src2 + offset)));
offset += HA;
v128_u16 hi = DivideBy16(BinomialSum(Load<align>(buffer.src0 + offset), Load<align>(buffer.src1 + offset), Load<align>(buffer.src2 + offset)));
return vec_pack(lo, hi);
}
template<bool align> SIMD_INLINE __m128i BlurRow16(const Buffer & buffer, size_t offset)
{
return DivideBy16(BinomialSum16(
Load<align>((__m128i*)(buffer.src0 + offset)),
Load<align>((__m128i*)(buffer.src1 + offset)),
Load<align>((__m128i*)(buffer.src2 + offset))));
}
