Half-Precision (16-bit) Float Point Numbers

Functions for conversion between 16-bit and 32-bit float numbers and other. More...

Functions
SIMD_API void	SimdFloat32ToFloat16 (const float *src, size_t size, uint16_t *dst)
	Converts numbers in the array from 32-bit float to 16-bit float format. More...
SIMD_API void	SimdFloat16ToFloat32 (const uint16_t *src, size_t size, float *dst)
	Converts numbers in the array from 16-bit float to 32-bit float format. More...
SIMD_API void	SimdSquaredDifferenceSum16f (const uint16_t *a, const uint16_t *b, size_t size, float *sum)
	Calculates sum of squared differences for two 16-bit float arrays. More...
SIMD_API void	SimdCosineDistance16f (const uint16_t *a, const uint16_t *b, size_t size, float *distance)
	Calculates cosine distance of two 16-bit float arrays. More...
SIMD_API void	SimdCosineDistancesMxNa16f (size_t M, size_t N, size_t K, const uint16_t *const *A, const uint16_t *const *B, float *distances)
	Calculates mutual cosine distance of two arrays of 16-bit float arrays. More...
SIMD_API void	SimdCosineDistancesMxNp16f (size_t M, size_t N, size_t K, const uint16_t *A, const uint16_t *B, float *distances)
	Calculates mutual cosine distance of two arrays of 16-bit float arrays. More...
SIMD_API void	SimdVectorNormNa16f (size_t N, size_t K, const uint16_t *const *A, float *norms)
	Calculates vector norms for array of 16-bit float arrays. More...
SIMD_API void	SimdVectorNormNp16f (size_t N, size_t K, const uint16_t *A, float *norms)
	Calculates vector norms for array of 16-bit float arrays. More...

Detailed Description

Functions for conversion between 16-bit and 32-bit float numbers and other.

Function Documentation

SimdFloat32ToFloat16()

void SimdFloat32ToFloat16	(	const float *	src,
		size_t	size,
		uint16_t *	dst
	)

Converts numbers in the array from 32-bit float to 16-bit float format.

Parameters

[in]	src	- a pointer to the input array with 32-bit float point numbers.
[in]	size	- a size of input and output array.
[out]	dst	- a pointer to the output array with 16-bit float point numbers.

SimdFloat16ToFloat32()

void SimdFloat16ToFloat32	(	const uint16_t *	src,
		size_t	size,
		float *	dst
	)

Converts numbers in the array from 16-bit float to 32-bit float format.

Parameters

[in]	src	- a pointer to the input array with 16-bit float point numbers.
[in]	size	- a size of input and output array.
[out]	dst	- a pointer to the output array with 32-bit float point numbers.

SimdSquaredDifferenceSum16f()

void SimdSquaredDifferenceSum16f	(	const uint16_t *	a,
		const uint16_t *	b,
		size_t	size,
		float *	sum
	)

Calculates sum of squared differences for two 16-bit float arrays.

All arrays must have the same size.

For every element:

sum += (a[i] - b[i])*(a[i] - b[i]);

Parameters

[in]	a	- a pointer to the first 16-bit float array.
[in]	b	- a pointer to the second 16-bit float array.
[in]	size	- a size of arrays.
[out]	sum	- a pointer to 32-bit float point sum of squared differences.

SimdCosineDistance16f()

void SimdCosineDistance16f	(	const uint16_t *	a,
		const uint16_t *	b,
		size_t	size,
		float *	distance
	)

Calculates cosine distance of two 16-bit float arrays.

All arrays must have the same size.

Algorithm description:

distance = 1 - Sum(a[i]*b[i])/Sqrt(Sum(a[i]*a[i])*Sum(b[i]*b[i]));

Parameters

[in]	a	- a pointer to the first 16-bit float array.
[in]	b	- a pointer to the second 16-bit float array.
[in]	size	- a size of arrays.
[out]	distance	- a pointer to 32-bit float with cosine distance.

SimdCosineDistancesMxNa16f()

void SimdCosineDistancesMxNa16f	(	size_t	M,
		size_t	N,
		size_t	K,
		const uint16_t *const *	A,
		const uint16_t *const *	B,
		float *	distances
	)

Calculates mutual cosine distance of two arrays of 16-bit float arrays.

Algorithm description:

distances[i, j] = 1 - Sum(A[i][k]*B[j][k])/Sqrt(Sum(A[i][k]*A[i][k])*Sum(B[j][k]*B[j][k]));

Parameters

[in]	M	- a number of A arrays.
[in]	N	- a number of B arrays.
[in]	K	- a size of A and B arrays.
[in]	A	- a pointer to the first array with pointers to 16-bit float arrays.
[in]	B	- a pointer to the second array with pointers to 16-bit float arrays.
[out]	distances	- a pointer to result 32-bit float array with cosine distances. It size must be M*N.

SimdCosineDistancesMxNp16f()

void SimdCosineDistancesMxNp16f	(	size_t	M,
		size_t	N,
		size_t	K,
		const uint16_t *	A,
		const uint16_t *	B,
		float *	distances
	)

Calculates mutual cosine distance of two arrays of 16-bit float arrays.

Algorithm description:

distances[i, j] = 1 - Sum(A[i*K + k]*B[j*K + k])/Sqrt(Sum(A[i*K + k]*A[i*K + k])*Sum(B[j*K + k]*B[j*K + k]));

Parameters

[in]	M	- a number of A arrays.
[in]	N	- a number of B arrays.
[in]	K	- a size of A and B arrays.
[in]	A	- a pointer to 16-bit float arrays.
[in]	B	- a pointer to 16-bit float arrays.
[out]	distances	- a pointer to result 32-bit float array with cosine distances. It size must be M*N.

SimdVectorNormNa16f()

void SimdVectorNormNa16f	(	size_t	N,
		size_t	K,
		const uint16_t *const *	A,
		float *	norms
	)

Calculates vector norms for array of 16-bit float arrays.

Algorithm description:

norms[j] = Sqrt(Sum(A[j][k]*A[j][k]));

Parameters

[in]	N	- a number of A arrays.
[in]	K	- a size of A arrays.
[in]	A	- a pointer to the array with pointers to 16-bit float arrays.
[out]	norms	- a pointer to result 32-bit float array with vector norms. It size must be N.

SimdVectorNormNp16f()

void SimdVectorNormNp16f	(	size_t	N,
		size_t	K,
		const uint16_t *	A,
		float *	norms
	)

Calculates vector norms for array of 16-bit float arrays.

Algorithm description:

norms[j] = Sqrt(Sum(A[j*K + k]*A[j*K + k]));

Parameters

[in]	N	- a number of A arrays.
[in]	K	- a size of A arrays.
[in]	A	- a pointer to 16-bit float arrays.
[out]	norms	- a pointer to result 32-bit float array with vector norms. It size must be N.