Moves between Scalar and Vector Unit

These intrinsics allow inserting or extracting of an individual element into/from a vector. More...

Overview

These intrinsics allow inserting or extracting of an individual element into/from a vector.

Inserting can be done either by pushing in a new value at one end of the vector and shifting or replacing a specific element.

Note

An int is converted to a int16 by taking the 16 LSB.

An int is converted to a cint16 by using the 16 LSB as the real part and the 16 MSB as the imaginary part.

There is no int8/uint8 support for these intrincs.

Update element in a vector with a value
v64int16	upd_elem (v64int16 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v32int16	upd_elem (v32int16 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v16int16	upd_elem (v16int16 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v8int16	upd_elem (v8int16 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v32int32	upd_elem (v32int32 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v16int32	upd_elem (v16int32 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v8int32	upd_elem (v8int32 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v4int32	upd_elem (v4int32 a, unsigned idx, int b)
	Update element determined by idx in vector a by value b.
v32cint16	upd_elem (v32cint16 a, unsigned idx, cint16 b)
	Update element determined by idx in vector a by value b.
v16cint16	upd_elem (v16cint16 a, unsigned idx, cint16 b)
	Update element determined by idx in vector a by value b.
v8cint16	upd_elem (v8cint16 a, unsigned idx, cint16 b)
	Update element determined by idx in vector a by value b.
v4cint16	upd_elem (v4cint16 a, unsigned idx, cint16 b)
	Update element determined by idx in vector a by value b.
v32float	upd_elem (v32float a, unsigned idx, float b)
	Update element determined by idx in vector a by value b.
v16float	upd_elem (v16float a, unsigned idx, float b)
	Update element determined by idx in vector a by value b.
v8float	upd_elem (v8float a, unsigned idx, float b)
	Update element determined by idx in vector a by value b.
v4float	upd_elem (v4float a, unsigned idx, float b)
	Update element determined by idx in vector a by value b.

Update elements in a vector with two 32 bit values
v32int32	upd_elem (v32int32 a, unsigned idx, int b0, int b1)
	Update elements determined by idx in vector a by values b0 and b1.
v16int32	upd_elem (v16int32 a, unsigned idx, int b0, int b1)
	Update elements determined by idx in vector a by values b0 and b1.
v8int32	upd_elem (v8int32 a, unsigned idx, int b0, int b1)
	Update elements determined by idx in vector a by values b0 and b1.
v4int32	upd_elem (v4int32 a, unsigned idx, int b0, int b1)
	Update elements determined by idx in vector a by values b0 and b1.
v32cint16	upd_elem (v32cint16 a, unsigned idx, cint16 b0, cint16 b1)
	Update elements determined by idx in vector a by values b0 and b1.
v16cint16	upd_elem (v16cint16 a, unsigned idx, cint16 b0, cint16 b1)
	Update elements determined by idx in vector a by values b0 and b1.
v8cint16	upd_elem (v8cint16 a, unsigned idx, cint16 b0, cint16 b1)
	Update elements determined by idx in vector a by values b0 and b1.
v4cint16	upd_elem (v4cint16 a, unsigned idx, cint16 b0, cint16 b1)
	Update elements determined by idx in vector a by values b0 and b1.
v32float	upd_elem (v32float a, unsigned idx, float b0, float b1)
	Update elements determined by idx in vector a by values b0 and b1.
v16float	upd_elem (v16float a, unsigned idx, float b0, float b1)
	Update elements determined by idx in vector a by values b0 and b1.
v8float	upd_elem (v8float a, unsigned idx, float b0, float b1)
	Update elements determined by idx in vector a by values b0 and b1.
v4float	upd_elem (v4float a, unsigned idx, float b0, float b1)
	Update elements determined by idx in vector a by values b0 and b1.

Update element in a vector with a complex 32-bit value
v16cint32	upd_elem (v16cint32 a, unsigned idx, cint32 b)
	Update element determined by idx in vector a by value b.
v8cint32	upd_elem (v8cint32 a, unsigned idx, cint32 b)
	Update element determined by idx in vector a by value b.
v4cint32	upd_elem (v4cint32 a, unsigned idx, cint32 b)
	Update element determined by idx in vector a by value b.
v2cint32	upd_elem (v2cint32 a, unsigned idx, cint32 b)
	Update element determined by idx in vector a by value b.
v16cfloat	upd_elem (v16cfloat a, unsigned idx, cfloat b)
	Update element determined by idx in vector a by value b.
v8cfloat	upd_elem (v8cfloat a, unsigned idx, cfloat b)
	Update element determined by idx in vector a by value b.
v4cfloat	upd_elem (v4cfloat a, unsigned idx, cfloat b)
	Update element determined by idx in vector a by value b.
v2cfloat	upd_elem (v2cfloat a, unsigned idx, cfloat b)
	Update element determined by idx in vector a by value b.

Extract an element from a vector
int	ext_elem (v64int16 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v32int16 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v16int16 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v8int16 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v32int32 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v16int32 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v8int32 a, unsigned idx)
	Extract element determined by idx from vector a.
int	ext_elem (v4int32 a, unsigned idx)
	Extract element determined by idx from vector a.
cint16	ext_elem (v32cint16 a, unsigned idx)
	Extract element determined by idx from vector a.
cint16	ext_elem (v16cint16 a, unsigned idx)
	Extract element determined by idx from vector a.
cint16	ext_elem (v8cint16 a, unsigned idx)
	Extract element determined by idx from vector a.
cint16	ext_elem (v4cint16 a, unsigned idx)
	Extract element determined by idx from vector a.
cint32	ext_elem (v16cint32 a, unsigned idx)
	Extract element determined by idx from vector a.
cint32	ext_elem (v8cint32 a, unsigned idx)
	Extract element determined by idx from vector a.
cint32	ext_elem (v4cint32 a, unsigned idx)
	Extract element determined by idx from vector a.
cint32	ext_elem (v2cint32 a, unsigned idx)
	Extract element determined by idx from vector a.
float	ext_elem (v32float a, unsigned idx)
	Extract element determined by idx from vector a.
float	ext_elem (v16float a, unsigned idx)
	Extract element determined by idx from vector a.
float	ext_elem (v8float a, unsigned idx)
	Extract element determined by idx from vector a.
float	ext_elem (v4float a, unsigned idx)
	Extract element determined by idx from vector a.
cfloat	ext_elem (v16cfloat a, unsigned idx)
	Extract element determined by idx from vector a.
cfloat	ext_elem (v8cfloat a, unsigned idx)
	Extract element determined by idx from vector a.
cfloat	ext_elem (v4cfloat a, unsigned idx)
	Extract element determined by idx from vector a.
cfloat	ext_elem (v2cfloat a, unsigned idx)
	Extract element determined by idx from vector a.

Extract a 20 bit element from a vector
addr_t	ext_addr (v32int32 a, unsigned idx)
	Extract element determined by idx from vector a.
addr_t	ext_addr (v16int32 a, unsigned idx)
	Extract element determined by idx from vector a.
addr_t	ext_addr (v8int32 a, unsigned idx)
	Extract element determined by idx from vector a.
addr_t	ext_addr (v4int32 a, unsigned idx)
	Extract element determined by idx from vector a.

Extract two 20 bit elements from a vector
void	ext_addr (v32int32 a, unsigned idx, addr_t &b0, addr_t &b1)
	Extract element determined by idx from vector a.
void	ext_addr (v16int32 a, unsigned idx, addr_t &b0, addr_t &b1)
	Extract element determined by idx from vector a.
void	ext_addr (v8int32 a, unsigned idx, addr_t &b0, addr_t &b1)
	Extract element determined by idx from vector a.
void	ext_addr (v4int32 a, unsigned idx, addr_t &b0, addr_t &b1)
	Extract element determined by idx from vector a.

Shift elements into a vector
v8int16	shft_elem (v8int16 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v4cint16	shft_elem (v4cint16 vec, cint16 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v4int32	shft_elem (v4int32 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v4float	shft_elem (v4float vec, float n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v2cint32	shft_elem (v2cint32 vec, cint32 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v2cfloat	shft_elem (v2cfloat vec, cfloat n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v16int16	shft_elem (v16int16 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v8cint16	shft_elem (v8cint16 vec, cint16 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v8int32	shft_elem (v8int32 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v8float	shft_elem (v8float vec, float n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v4cint32	shft_elem (v4cint32 vec, cint32 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v4cfloat	shft_elem (v4cfloat vec, cfloat n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v32int16	shft_elem (v32int16 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v16cint16	shft_elem (v16cint16 vec, cint16 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v16int32	shft_elem (v16int32 vec, int n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v16float	shft_elem (v16float vec, float n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v8cint32	shft_elem (v8cint32 vec, cint32 n)
	Insert a new element at the beginning of a vector by shifting the other elements.
v8cfloat	shft_elem (v8cfloat vec, cfloat n)
	Insert a new element at the beginning of a vector by shifting the other elements.

Function Documentation

ext_addr() [1/8]

addr_t ext_addr	(	v16int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

Returns

unsigned 20 bit (lower bits) from the corresponding 32 bit vector element

ext_addr() [2/8]

void ext_addr	(	v16int32	a,
		unsigned	idx,
		addr_t &	b0,
		addr_t &	b1
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector
b0	20 bit value
b1	20 bit value

Extracts 2 20 bit values from the vector and places the lower indexed (2*idx) element into b0 and the higher indexed (2*idx + 1) into b1. The lower 20 bits of a 32 bit vector element are stored in b0/b1.

ext_addr() [3/8]

addr_t ext_addr	(	v32int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

Returns

unsigned 20 bit (lower bits) from the corresponding 32 bit vector element

ext_addr() [4/8]

void ext_addr	(	v32int32	a,
		unsigned	idx,
		addr_t &	b0,
		addr_t &	b1
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector
b0	20 bit value
b1	20 bit value

Extracts 2 20 bit values from the vector and places the lower indexed (2*idx) element into b0 and the higher indexed (2*idx + 1) into b1. The lower 20 bits of a 32 bit vector element are stored in b0/b1.

ext_addr() [5/8]

addr_t ext_addr	(	v4int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

Returns

unsigned 20 bit (lower bits) from the corresponding 32 bit vector element

ext_addr() [6/8]

void ext_addr	(	v4int32	a,
		unsigned	idx,
		addr_t &	b0,
		addr_t &	b1
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector
b0	20 bit value
b1	20 bit value

Extracts 2 20 bit values from the vector and places the lower indexed (2*idx) element into b0 and the higher indexed (2*idx + 1) into b1. The lower 20 bits of a 32 bit vector element are stored in b0/b1.

ext_addr() [7/8]

addr_t ext_addr	(	v8int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

Returns

unsigned 20 bit (lower bits) from the corresponding 32 bit vector element

ext_addr() [8/8]

void ext_addr	(	v8int32	a,
		unsigned	idx,
		addr_t &	b0,
		addr_t &	b1
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector
b0	20 bit value
b1	20 bit value

Extracts 2 20 bit values from the vector and places the lower indexed (2*idx) element into b0 and the higher indexed (2*idx + 1) into b1. The lower 20 bits of a 32 bit vector element are stored in b0/b1.

ext_elem() [1/24]

cfloat ext_elem	(	v16cfloat	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [2/24]

cint16 ext_elem	(	v16cint16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [3/24]

cint32 ext_elem	(	v16cint32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [4/24]

float ext_elem	(	v16float	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [5/24]

int ext_elem	(	v16int16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [6/24]

int ext_elem	(	v16int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [7/24]

cfloat ext_elem	(	v2cfloat	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [8/24]

cint32 ext_elem	(	v2cint32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [9/24]

cint16 ext_elem	(	v32cint16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [10/24]

float ext_elem	(	v32float	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [11/24]

int ext_elem	(	v32int16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [12/24]

int ext_elem	(	v32int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [13/24]

cfloat ext_elem	(	v4cfloat	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [14/24]

cint16 ext_elem	(	v4cint16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [15/24]

cint32 ext_elem	(	v4cint32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [16/24]

float ext_elem	(	v4float	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [17/24]

int ext_elem	(	v4int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [18/24]

int ext_elem	(	v64int16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [19/24]

cfloat ext_elem	(	v8cfloat	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [20/24]

cint16 ext_elem	(	v8cint16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [21/24]

cint32 ext_elem	(	v8cint32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [22/24]

float ext_elem	(	v8float	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [23/24]

int ext_elem	(	v8int16	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

ext_elem() [24/24]

int ext_elem	(	v8int32	a,
		unsigned	idx
	)

Extract element determined by idx from vector a.

Parameters

a	The vector from which the element will be extracted from
idx	Index used to choose element in vector

shft_elem() [1/18]

v16cint16 shft_elem	(	v16cint16	vec,
		cint16	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [2/18]

v16float shft_elem	(	v16float	vec,
		float	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [3/18]

v16int16 shft_elem	(	v16int16	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [4/18]

v16int32 shft_elem	(	v16int32	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [5/18]

v2cfloat shft_elem	(	v2cfloat	vec,
		cfloat	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [6/18]

v2cint32 shft_elem	(	v2cint32	vec,
		cint32	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [7/18]

v32int16 shft_elem	(	v32int16	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [8/18]

v4cfloat shft_elem	(	v4cfloat	vec,
		cfloat	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [9/18]

v4cint16 shft_elem	(	v4cint16	vec,
		cint16	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [10/18]

v4cint32 shft_elem	(	v4cint32	vec,
		cint32	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [11/18]

v4float shft_elem	(	v4float	vec,
		float	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [12/18]

v4int32 shft_elem	(	v4int32	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [13/18]

v8cfloat shft_elem	(	v8cfloat	vec,
		cfloat	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [14/18]

v8cint16 shft_elem	(	v8cint16	vec,
		cint16	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [15/18]

v8cint32 shft_elem	(	v8cint32	vec,
		cint32	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [16/18]

v8float shft_elem	(	v8float	vec,
		float	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [17/18]

v8int16 shft_elem	(	v8int16	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

shft_elem() [18/18]

v8int32 shft_elem	(	v8int32	vec,
		int	n
	)

Insert a new element at the beginning of a vector by shifting the other elements.

These intrinsics shift all elements of a vector up by one element, therefore losing the highest element and insert a new element in the first position and then returns the vector. The following code demonstrates this operation:

for (i=<vector-size>-1; i>0; i--)
{
    vec[i]=vec[i-1];
}
vec[0] = n;

Returns

Shifted vector

Parameters

vec	Vector to be shifted and into which the new value is inserted.
n	New value that will be inserted into the vector.

DPD considerations

Within the DPD, this intrinsic is used to feed the incoming streams into the corresponding vectors. For example, it might be used like this:

int frac0;
split(get_ss0(), mag_scale, lut_width, idx[0], frac0);
frac = shft_elem(frac, delay2(frac0));
data = shft_elem(data, get_ss1());

upd_elem() [1/36]

v16cfloat upd_elem	(	v16cfloat	a,
		unsigned	idx,
		cfloat	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [2/36]

v16cint16 upd_elem	(	v16cint16	a,
		unsigned	idx,
		cint16	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [3/36]

v16cint16 upd_elem	(	v16cint16	a,
		unsigned	idx,
		cint16	b0,
		cint16	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [4/36]

v16cint32 upd_elem	(	v16cint32	a,
		unsigned	idx,
		cint32	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [5/36]

v16float upd_elem	(	v16float	a,
		unsigned	idx,
		float	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [6/36]

v16float upd_elem	(	v16float	a,
		unsigned	idx,
		float	b0,
		float	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [7/36]

v16int16 upd_elem	(	v16int16	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [8/36]

v16int32 upd_elem	(	v16int32	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [9/36]

v16int32 upd_elem	(	v16int32	a,
		unsigned	idx,
		int	b0,
		int	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [10/36]

v2cfloat upd_elem	(	v2cfloat	a,
		unsigned	idx,
		cfloat	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [11/36]

v2cint32 upd_elem	(	v2cint32	a,
		unsigned	idx,
		cint32	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [12/36]

v32cint16 upd_elem	(	v32cint16	a,
		unsigned	idx,
		cint16	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [13/36]

v32cint16 upd_elem	(	v32cint16	a,
		unsigned	idx,
		cint16	b0,
		cint16	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [14/36]

v32float upd_elem	(	v32float	a,
		unsigned	idx,
		float	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [15/36]

v32float upd_elem	(	v32float	a,
		unsigned	idx,
		float	b0,
		float	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [16/36]

v32int16 upd_elem	(	v32int16	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [17/36]

v32int32 upd_elem	(	v32int32	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [18/36]

v32int32 upd_elem	(	v32int32	a,
		unsigned	idx,
		int	b0,
		int	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [19/36]

v4cfloat upd_elem	(	v4cfloat	a,
		unsigned	idx,
		cfloat	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [20/36]

v4cint16 upd_elem	(	v4cint16	a,
		unsigned	idx,
		cint16	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [21/36]

v4cint16 upd_elem	(	v4cint16	a,
		unsigned	idx,
		cint16	b0,
		cint16	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [22/36]

v4cint32 upd_elem	(	v4cint32	a,
		unsigned	idx,
		cint32	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [23/36]

v4float upd_elem	(	v4float	a,
		unsigned	idx,
		float	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [24/36]

v4float upd_elem	(	v4float	a,
		unsigned	idx,
		float	b0,
		float	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [25/36]

v4int32 upd_elem	(	v4int32	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [26/36]

v4int32 upd_elem	(	v4int32	a,
		unsigned	idx,
		int	b0,
		int	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [27/36]

v64int16 upd_elem	(	v64int16	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [28/36]

v8cfloat upd_elem	(	v8cfloat	a,
		unsigned	idx,
		cfloat	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [29/36]

v8cint16 upd_elem	(	v8cint16	a,
		unsigned	idx,
		cint16	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [30/36]

v8cint16 upd_elem	(	v8cint16	a,
		unsigned	idx,
		cint16	b0,
		cint16	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [31/36]

v8cint32 upd_elem	(	v8cint32	a,
		unsigned	idx,
		cint32	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [32/36]

v8float upd_elem	(	v8float	a,
		unsigned	idx,
		float	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [33/36]

v8float upd_elem	(	v8float	a,
		unsigned	idx,
		float	b0,
		float	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.

upd_elem() [34/36]

v8int16 upd_elem	(	v8int16	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [35/36]

v8int32 upd_elem	(	v8int32	a,
		unsigned	idx,
		int	b
	)

Update element determined by idx in vector a by value b.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b	value to update.

upd_elem() [36/36]

v8int32 upd_elem	(	v8int32	a,
		unsigned	idx,
		int	b0,
		int	b1
	)

Update elements determined by idx in vector a by values b0 and b1.

Parameters

a	The vector containing the element to be updated
idx	Index used to choose where to update in vector a
b0	32 bit value
b1	32 bit value

32 bits from both b0 and b1 are used to update two consecutive 32 bit elements of vector a. The elements are grouped into pairs such as 0+1, 2+3, etc. and idx is used to select which pair is updated. b0 goes in the lower bits (idx*2) and b1 goes in the higher bits (idx*2 + 1). An idx of 5 would select the 5th pairing which would be elements 10 and 11.