nrf_axon_axpb_2d_8_16()

nrf_axon_result_e nrf_axon_axpb_2d_8_16	(	const int8_t *	x_ptr,
		int32_t	a_scalar,
		int32_t	b_scalar,
		int16_t *	output,
		uint16_t	height,
		uint16_t	width,
		bool	input_is_packed,
		uint8_t	rounding_bits,
		nrf_axon_syncmode_blocking_e	block_mode,
		bool	keep_reservation )

#include <include/drivers/axon/nrf_axon_dsp_intrinsics.h>

"axpb"=>"a times x + b" Multiplies each vector value by "a" and adds "b" Input vector can be 2 dimensional for improved performance. Note that if width is an odd number, each row of output will have a padding short at the end.

Parameters

x_ptr	x vector input. 8bit integers.
a_scalar	"a" scalar value to multiply with input , 32bit integer.
b_scalar	"b" scalar value to add to input, 32bit integer.
out_ptr	output vector. 16bit integers, no extra stride, and rounded by rounding_bits. output can overlap input.
height	Height of the input/output in rows. Should be less than width for optimal performance.
width	Width of the input/output in elements. Should be greater than height for optimal performance. If not a muultiple of 2, output width will be 0 padded rounded to the next multiple of 2. Maximum 1024.
bool	input_is_packed if true, input has no padding between rows. if false, input rows start on 32bit boundaries (n/a when width is a multiple of 4)
rounding_bits	Number of bits to round the output by. 0 => no rounding, maximum of 32.
block_mode	recommnended to be set to NRF_AXON_SYNC_MODE_BLOCKING_POLLING.
keep_reservation	set to true if there are subsequent axon operations to execute immediately on completion. true prevents axon from be taken by another user and/or being powered off. Set to false if axon will not be used immediately after to free axon for other users and/or power down if idle.

Returns

kAxonResultSuccess on success or a negative error code (see nrf_axon_result_e)