Upshift

Intrinsics for moving values from vector data-types to accumulator data-types. More...

Overview

Intrinsics for moving values from vector data-types to accumulator data-types.

Moving data from vector data-types to accumulator data-types (e.g. for initialization) requires adjustment in precision because the accumulator data-types are wider in size. For fixed point arithmetic, an appropriate left shift operation would align the decimal point between the two representations. The shift amount is specified as a parameter (in the range -1 to 62 which will be encoded as 0..63 in the instruction).

UPS/LUPS integer intrinsic functions

There are two variants of the UPS intrinsics based on width of input and output data-types:

1. ups is used to convert integer
   • 8-bit vector data into corresponding 48-bit accumulator, or
   • 16-bit vector data (short) into corresponding 48-bit accumulator, or
   • 32-bit vector data into corresponding 48-bit accumulator
2. lups is used to convert integer
   • 32-bit vector data into corresponding 80-bit accumulator, or
   • 64-bit vector data (long) into corresponding 80-bit accumulator

See also

'srs' Intrinsics (Shift-Round-Saturate)

256-bit vector to 768-bit accumulator conversions
v16acc48	ups (v16int16 a, int shft)
v8cacc48	ups (v8cint16 a, int shft)
v8acc80	lups (v8int32 a, int shft)
v4cacc80	lups (v4cint32 a, int shft)

256-bit vector to 384-bit accumulator conversions
v8acc48	ups (v8int32 a, int shft)
v4cacc48	ups (v4cint32 a, int shft)
v4acc80	lups (v4int64 a, int shft)
v2cacc80	lups (v2cint64 a, int shft)

128-bit vector to 384-bit accumulator conversions
v8acc48	ups (v8int16 a, int shft)
v4cacc48	ups (v4cint16 a, int shft)
v4acc80	lups (v4int32 a, int shft)
v2cacc80	lups (v2cint32 a, int shft)

128-bit vector to 768-bit accumulator conversions
v16acc48	ups (v16int8 a, int shft)
v16acc48	ups (v16uint8 a, int shft)

Function Documentation

lups() [1/6]

v2cacc80 lups	(	v2cint32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

lups() [2/6]

v2cacc80 lups	(	v2cint64	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

lups() [3/6]

v4cacc80 lups	(	v4cint32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

lups() [4/6]

v4acc80 lups	(	v4int32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

lups() [5/6]

v4acc80 lups	(	v4int64	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

lups() [6/6]

v8acc80 lups	(	v8int32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [1/8]

v16acc48 ups	(	v16int16	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [2/8]

v16acc48 ups	(	v16int8	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [3/8]

v16acc48 ups	(	v16uint8	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [4/8]

v4cacc48 ups	(	v4cint16	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [5/8]

v4cacc48 ups	(	v4cint32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [6/8]

v8cacc48 ups	(	v8cint16	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [7/8]

v8acc48 ups	(	v8int16	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.

ups() [8/8]

v8acc48 ups	(	v8int32	a,
		int	shft
	)

shft parameter is in the range of -1 to 62 which will encoded as 0..63 in the instruction.