maths.h File Reference

Support functions for STDP. More...

#include <common/neuron-typedefs.h>
#include <debug.h>
#include <spin1_api.h>

Go to the source code of this file.

Data Structures
struct	int16_lut
	Lookup Table of 16-bit integers. More...

Macros
#define	MIN(X, Y)   ((X) < (Y) ? (X) : (Y))
	Minimum. Evaluates arguments twice.
#define	MAX(X, Y)   ((X) > (Y) ? (X) : (Y))
	Maximum. Evaluates arguments twice.

Functions
static int16_lut *	maths_copy_int16_lut (address_t *address)
	Copy a Lookup Table from SDRAM to DTCM, updating the address.
static int32_t	maths_lut_exponential_decay (uint32_t time, const int16_lut *lut)
	Get value from lookup table.
static int32_t	maths_clamp_pot (int32_t x, uint32_t shift)
	Clamp to fit in number of bits.
static int32_t	maths_mul_16x16 (int16_t x, int16_t y)
	multiply two 16-bit numbers to get a 32-bit number.
static int32_t	maths_fixed_mul16 (int32_t a, int32_t b, const int32_t fixed_point_position)
	multiply two 16-bit fixed point numbers (encoded in int32_t)
static int32_t	maths_fixed_mul32 (int32_t a, int32_t b, const int32_t fixed_point_position)
	multiply two 32-bit fixed point numbers (encoded in int32_t)

Detailed Description

Support functions for STDP.

Definition in file maths.h.

---

Data Structure Documentation

int16_lut

struct int16_lut

Lookup Table of 16-bit integers.

Will be padded to a word boundary at the end.

Definition at line 44 of file maths.h.

Data Fields
uint16_t	size	Number of entries in table.
uint16_t	shift	Mapping from time to table index.
int16_t	values[]	Table of actual values.

Macro Definition Documentation

MIN

#define MIN	(		X,
			Y
	)		((X) < (Y) ? (X) : (Y))

Minimum. Evaluates arguments twice.

Parameters

X	First value
Y	Second value

Returns

Minimum of two values

Definition at line 34 of file maths.h.

MAX

#define MAX	(		X,
			Y
	)		((X) > (Y) ? (X) : (Y))

Maximum. Evaluates arguments twice.

Parameters

X	First value
Y	Second value

Returns

Maximum of two values

Definition at line 39 of file maths.h.

Function Documentation

maths_copy_int16_lut()

static int16_lut * maths_copy_int16_lut ( address_t *  address )

inlinestatic

Copy a Lookup Table from SDRAM to DTCM, updating the address.

Parameters

[in,out]

address

Pointer to the SDRAM address to copy from. This is updated to point to the space after the structure.

Returns

A pointer to the copied lookup table, malloc'd in DTCM

Definition at line 57 of file maths.h.

maths_lut_exponential_decay()

static int32_t maths_lut_exponential_decay ( uint32_t  time, const int16_lut *  lut  )

inlinestatic

Get value from lookup table.

Parameters

[in]	time	The time that we are mapping
[in]	lut	The lookup table (result of maths_copy_int16_lut())

Returns

The value from the LUT, or zero if out of range

Definition at line 79 of file maths.h.

maths_clamp_pot()

static int32_t maths_clamp_pot ( int32_t  x, uint32_t  shift  )

inlinestatic

Clamp to fit in number of bits.

Parameters

[in]	x	The value to clamp
[in]	shift	Width of the field to clamp the value to fit in

Returns

The clamped value

Definition at line 92 of file maths.h.

maths_mul_16x16()

static int32_t maths_mul_16x16 ( int16_t  x, int16_t  y  )

inlinestatic

multiply two 16-bit numbers to get a 32-bit number.

NOTE: this should 'encourage' GCC to insert SMULxy 16x16 multiply

Parameters

[in]	x	The first multiplicand
[in]	y	The first multiplicand

Returns

The product

Definition at line 109 of file maths.h.

maths_fixed_mul16()

static int32_t maths_fixed_mul16 ( int32_t  a, int32_t  b, const int32_t  fixed_point_position  )

inlinestatic

multiply two 16-bit fixed point numbers (encoded in int32_t)

Parameters

[in]	a	The first multiplicand
[in]	b	The first multiplicand
[in]	fixed_point_position	The location of the fixed point

Returns

The product

Definition at line 119 of file maths.h.

maths_fixed_mul32()

static int32_t maths_fixed_mul32 ( int32_t  a, int32_t  b, const int32_t  fixed_point_position  )

inlinestatic

multiply two 32-bit fixed point numbers (encoded in int32_t)

Parameters

[in]	a	The first multiplicand
[in]	b	The first multiplicand
[in]	fixed_point_position	The location of the fixed point

Returns

The product

Definition at line 131 of file maths.h.