Cache-friendly deinterleave RGB image - Comparison between standard and deinterleave for blocks

deinterleave_RGB_blocks.cpp

// DEINTERLEAVE RGB Image

template<typename T, typename U>
inline void deinterleave_BGR(const T* const interleaved_BGR, U** const deinterleaved_BGR, const uint32_t nsize) {

	constexpr uint32_t block = 262144 / (3 * std::max(sizeof(T), sizeof(U)));
	constexpr float round = std::is_integral_v<U> ? std::is_integral_v<T> ? 0 : 0.5f : 0;

#pragma omp parallel for
	for (int32_t x = 0; x < nsize; x += block)
	{
		U* const B = deinterleaved_BGR[0] + x;
		U* const G = deinterleaved_BGR[1] + x;
		U* const R = deinterleaved_BGR[2] + x;
		const T* const interleaved_ptr = interleaved_BGR + x * 3;

		const int blockx = std::min(nsize, x + block) - x;
		for (int xx = 0; xx < blockx; ++xx)
		{
			B[xx] = interleaved_ptr[xx * 3 + 0] + round;
			G[xx] = interleaved_ptr[xx * 3 + 1] + round;
			R[xx] = interleaved_ptr[xx * 3 + 2] + round;
		}
	}

}

template<typename T, typename U>
void interleave_BGR(const U** const deinterleaved_BGR, T* const interleaved_BGR, const uint32_t nsize) {

	constexpr uint32_t block = 262144 / (3 * std::max(sizeof(T), sizeof(U)));
	constexpr float round = std::is_integral_v<T> ? std::is_integral_v<U> ? 0 : 0.5f : 0;

#pragma omp parallel for
	for (int32_t x = 0; x < nsize; x += block)
	{
		const U* const B = deinterleaved_BGR[0] + x;
		const U* const G = deinterleaved_BGR[1] + x;
		const U* const R = deinterleaved_BGR[2] + x;
		T* const interleaved_ptr = interleaved_BGR + x * 3;

		const int blockx = std::min(nsize, x + block) - x;
		for (int xx = 0; xx < blockx; ++xx)
		{
			interleaved_ptr[xx * 3 + 0] = B[xx] + round;
			interleaved_ptr[xx * 3 + 1] = G[xx] + round;
			interleaved_ptr[xx * 3 + 2] = R[xx] + round;
		}
	}

}

Looks good.

Please note that I have already performed some implicit channel splits in my code to date. One can read in data in, in one format do processing as required, and then put it back in a different format subsequently. I have tended to keep the data in a single array. So instead of pointers initialised at 0, 0, and 0 for three different arrays, I have gone for pointers initialised at 0, green offset and blue offset for the same array.