SQRSHRNT

Signed saturating rounding shift right narrow by immediate (top)

Shift each signed integer value in the source vector elements right by an immediate value, and place the rounded results in the odd-numbered half-width destination elements, leaving the even-numbered elements unchanged. Each result element is saturated to the half-width N-bit element's signed integer range -2(N-1) to (2(N-1))-1. The immediate shift amount is an unsigned value in the range 1 to number of bits per element. This instruction is unpredicated.

Encoding: SVE2

Variants: FEAT_SVE2 || FEAT_SME (FEAT_SVE2 || FEAT_SME)

31	30	29	28	27	26	25	24	23	22	21	20	19	18	17	16	15	14	13	12	11	10	9	8	7	6	5	4	3	2	1	0
0	1	0	0	0	1	0	1	0		1						0	0	1	0	1	1
									tszh		tszl		imm3					op	U	R	T	Zn					Zd

SQRSHRNT <Zd>.<T>, <Zn>.<Tb>, #<const>

Decoding algorithm

if !IsFeatureImplemented(FEAT_SVE2) && !IsFeatureImplemented(FEAT_SME) then
    EndOfDecode(Decode_UNDEF);
constant bits(3) tsize = tszh:tszl;
if tsize == '000' then EndOfDecode(Decode_UNDEF);
constant integer esize = 8 << HighestSetBit(tsize);
constant integer n = UInt(Zn);
constant integer d = UInt(Zd);
constant integer shift = (2 * esize) - UInt(tsize:imm3);

Operation

CheckSVEEnabled();
constant integer VL = CurrentVL;
constant integer elements = VL DIV (2 * esize);
constant bits(VL) operand = Z[n, VL];
bits(VL) result = Z[d, VL];
for e = 0 to elements-1
    constant bits(2*esize) element = Elem[operand, e, 2*esize];
    constant integer res = (SInt(element) + (1 << (shift-1))) >> shift;
    Elem[result, 2*e + 1, esize] = SignedSat(res, esize);

Z[d, VL] = result;

Explanations

<Zd>: Is the name of the destination scalable vector register, encoded in the "Zd" field.
<T>: <Zn>: Is the name of the source scalable vector register, encoded in the "Zn" field.
<Tb>: <const>: Is the immediate shift amount, in the range 1 to number of bits per element, encoded in "tszh:tszl:imm3".