findIntSSE2

findIntSSE2

const int* findIntSSE2(const int* __restrict begin, const int* __restrict end, int needle)
{
	const int* it = begin;
	unsigned long index;
	__m128i n = _mm_set1_epi32(needle);
	while (it != end)
	{		
		__m128i x = _mm_load_si128(reinterpret_cast<const __m128i*>(it));
		__m128i m = _mm_cmpeq_epi32(x, n);
		unsigned long k = _mm_movemask_epi8(m);
		if (_BitScanForward(&index, k))
		{
			return it + index/4;
		}
		it += 4;
	}
	return it;
}