Overview

Advanced Float Vector Arithmetic Operations. The lane selection scheme is explained after each intrinsic. Note that these intrinsics are a wrapper over the fully configurable fpmac_conf base function, explained here. More information on fpmac_conf can be found here.

Functions
v8float	fpabs (v32float xbuf, int xstart, unsigned int xoffs)
	Select elements and take absolute value of elements from a real single precision floating point vector.

v8float	fpabs (v16float xbuf, int xstart, unsigned int xoffs)
	Select elements and take absolute value of elements from a real single precision floating point vector.

Addition functions
v8float	fpadd (v8float acc, v32float xbuf, int xstart, unsigned int xoffs)
	Add single precision real floating point vectors.

v8float	fpadd (v8float acc, v16float xbuf, int xstart, unsigned int xoffs)
	Add single precision real floating point vectors.

v4cfloat	fpadd (v4cfloat acc, v16cfloat xbuf, int xstart, unsigned int xoffs)
	Add complex single precision floating point vectors.

v4cfloat	fpadd (v4cfloat acc, v8cfloat xbuf, int xstart, unsigned int xoffs)
	Add complex single precision floating point vectors.

v8float	fpadd_abs (v8float acc, v32float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and add real single precision floating point vectors.

v8float	fpadd_abs (v8float acc, v16float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and add real single precision floating point vectors.

Subtraction functions
v8float	fpsub (v8float acc, v32float xbuf, int xstart, unsigned int xoffs)
	Subtract real single precision floating point vectors.

v8float	fpsub (v8float acc, v16float xbuf, int xstart, unsigned int xoffs)
	Subtract real single precision floating point vectors.

v4cfloat	fpsub (v4cfloat acc, v16cfloat xbuf, int xstart, unsigned int xoffs)
	Subtract complex single precision floating point vectors.

v4cfloat	fpsub (v4cfloat acc, v8cfloat xbuf, int xstart, unsigned int xoffs)
	Subtract complex single precision floating point vectors.

v8float	fpsub_abs (v8float acc, v32float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and subtract real single precision floating point vectors.

v8float	fpsub_abs (v8float acc, v16float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and subtract real single precision floating point vectors.

Negation intrinsics
v8float	fpneg (v32float xbuf, int xstart, unsigned int xoffs)
	Negate a real single precision floating point vector.

v8float	fpneg (v16float xbuf, int xstart, unsigned int xoffs)
	Negate a real single precision floating point vector.

v4cfloat	fpneg (v16cfloat xbuf, int xstart, unsigned int xoffs)
	Negate a complex single precision floating point vector.

v4cfloat	fpneg (v8cfloat xbuf, int xstart, unsigned int xoffs)
	Negate a complex single precision floating point vector.

v8float	fpneg_abs (v32float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and negate a real single precision floating point vector.

v8float	fpneg_abs (v16float xbuf, int xstart, unsigned int xoffs)
	Take absolute value and negate a real single precision floating point vector.

Function Documentation

v8float fpabs	(	v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Select elements and take absolute value of elements from a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the absolute values.

Parameters

xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpabs	(	v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Select elements and take absolute value of elements from a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the absolute values.

Parameters

xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpadd	(	v8float	acc,
		v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Add single precision real floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] + xbuf[xstart + xoffs[i]]

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Big input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpadd	(	v8float	acc,
		v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Add single precision real floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] + xbuf[xstart + xoffs[i]]

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v4cfloat fpadd	(	v4cfloat	acc,
		v16cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Add complex single precision floating point vectors.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = acc[i].re + xbuf[xstart + xoffs[i]].re
  ret[i].re = acc[i].im + xbuf[xstart + xoffs[i]].im

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v4cfloat fpadd	(	v4cfloat	acc,
		v8cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Add complex single precision floating point vectors.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = acc[i].re + xbuf[xstart + xoffs[i]].re
  ret[i].re = acc[i].im + xbuf[xstart + xoffs[i]].im

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v8float fpadd_abs	(	v8float	acc,
		v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and add real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] + abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpadd_abs	(	v8float	acc,
		v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and add real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] + abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the addition result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpneg	(	v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Negate a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = - xbuf[xstart + xoffs[i]]

Returns: Vector with the negated result.

Parameters

xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpneg	(	v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Negate a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = - xbuf[xstart + xoffs[i]]

Returns: Vector with the negated result.

Parameters

xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v4cfloat fpneg	(	v16cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Negate a complex single precision floating point vector.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = - xbuf[xstart + xoffs[i]].re
  ret[i].im = - xbuf[xstart + xoffs[i]].im

Returns: Vector with the subtraction result.

Parameters

xbuf	Input buffer.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v4cfloat fpneg	(	v8cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Negate a complex single precision floating point vector.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = - xbuf[xstart + xoffs[i]].re
  ret[i].im = - xbuf[xstart + xoffs[i]].im

Returns: Vector with the subtraction result.

Parameters

xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v8float fpneg_abs	(	v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and negate a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = - abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the negated absolute values.

Parameters

xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpneg_abs	(	v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and negate a real single precision floating point vector.

for (i = 0 ; i < 8 ; i++)

ret[i] = - abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the negated absolute values.

Parameters

xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpsub	(	v8float	acc,
		v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Subtract real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] - xbuf[xstart + xoffs[i]]

Returns: Vector with the subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpsub	(	v8float	acc,
		v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Subtract real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] - xbuf[xstart + xoffs[i]]

Returns: Vector with the subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v4cfloat fpsub	(	v4cfloat	acc,
		v16cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Subtract complex single precision floating point vectors.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = acc[i].re - xbuf[xstart + xoffs[i]].re
  ret[i].re = acc[i].im - xbuf[xstart + xoffs[i]].im

Returns: Vector with the subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v4cfloat fpsub	(	v4cfloat	acc,
		v8cfloat	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Subtract complex single precision floating point vectors.

for (i = 0 ; i < 4 ; i++)
  ret[i].re = acc[i].re - xbuf[xstart + xoffs[i]].re
  ret[i].re = acc[i].im - xbuf[xstart + xoffs[i]].im

Returns: Vector with the subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X. The offsets are referring to complex lanes (lane 0 corresponds to the first real and complex values).
xoffs	4 bits per lane: Additional lane-dependent offset for X. The offsets are referring to complex lanes. For optimized code should be compile time constant.

v8float fpsub_abs	(	v8float	acc,
		v32float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and subtract real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] - abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the abs-subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

v8float fpsub_abs	(	v8float	acc,
		v16float	xbuf,
		int	xstart,
		unsigned int	xoffs
	)

Take absolute value and subtract real single precision floating point vectors.

for (i = 0 ; i < 8 ; i++)

ret[i] = acc[i] - abs(xbuf[xstart + xoffs[i]])

Returns: Vector with the abs-subtraction result.

Parameters

acc	Incoming accumulation vector.
xbuf	Input buffer. Small buffer variant.
xstart	Starting offset for all lanes of X.
xoffs	4 bits per lane: Additional lane-dependent offset for X.

Overview

Functions

Addition functions

Subtraction functions

Negation intrinsics

Function Documentation