Basic Math Functions for Neural Network Computation

Functions
void	arm_nn_accumulate_q7_to_q15 (q15_t *pDst, const q7_t *pSrc, uint32_t length)
	Converts the elements from a q7 vector and accumulate to a q15 vector. More...
void	arm_nn_add_q7 (const q7_t *input, q31_t *output, uint32_t block_size)
	Non-saturating addition of elements of a q7 vector. More...
void	arm_nn_mult_q15 (q15_t *pSrcA, q15_t *pSrcB, q15_t *pDst, const uint16_t out_shift, uint32_t blockSize)
	Q7 vector multiplication with variable output shifts. More...
void	arm_nn_mult_q7 (q7_t *pSrcA, q7_t *pSrcB, q7_t *pDst, const uint16_t out_shift, uint32_t blockSize)
	Q7 vector multiplication with variable output shifts. More...

Description

Basic Math Functions for Neural Network Computation

Function Documentation

void arm_nn_accumulate_q7_to_q15	(	q15_t *	dst,
		const q7_t *	src,
		uint32_t	block_size
	)

Parameters

[in]	*src	points to the q7 input vector
[out]	*dst	points to the q15 output vector
[in]	block_size	length of the input vector

Description:

The equation used for the conversion process is:

 dst[n] += (q15_t) src[n] ;   0 <= n < block_size.

References arm_nn_read_q15x2(), and arm_nn_read_q7x4_ia().

Referenced by arm_avgpool_s8().

void arm_nn_add_q7	(	const q7_t *	input,
		q31_t *	output,
		uint32_t	block_size
	)

Parameters

[in]	*input	Pointer to the q7 input vector
[out]	*output	Pointer to the q31 output variable.
[in]	block_size	length of the input vector

Description:

2^24 samples can be added without saturating the result.

The equation used for the conversion process is:

 sum = input[0] + input[1] + .. + input[block_size -1]

References arm_nn_read_q7x4_ia().

void arm_nn_mult_q15	(	q15_t *	pSrcA,
		q15_t *	pSrcB,
		q15_t *	pDst,
		const uint16_t	out_shift,
		uint32_t	blockSize
	)

q7 vector multiplication with variable output shifts

Parameters

[in]	*pSrcA	pointer to the first input vector
[in]	*pSrcB	pointer to the second input vector
[out]	*pDst	pointer to the output vector
[in]	out_shift	amount of right-shift for output
[in]	blockSize	number of samples in each vector

Scaling and Overflow Behavior:

The function uses saturating arithmetic. Results outside of the allowable Q15 range [0x8000 0x7FFF] will be saturated.

References NN_ROUND.

void arm_nn_mult_q7	(	q7_t *	pSrcA,
		q7_t *	pSrcB,
		q7_t *	pDst,
		const uint16_t	out_shift,
		uint32_t	blockSize
	)

q7 vector multiplication with variable output shifts

Parameters

[in]	*pSrcA	pointer to the first input vector
[in]	*pSrcB	pointer to the second input vector
[out]	*pDst	pointer to the output vector
[in]	out_shift	amount of right-shift for output
[in]	blockSize	number of samples in each vector

Scaling and Overflow Behavior:

The function uses saturating arithmetic. Results outside of the allowable Q7 range [0x80 0x7F] will be saturated.

References NN_ROUND.