SQRSHRN

Multi-vector signed saturating rounding shift right narrow by immediate and interleave

This instruction shifts right by an immediate value the signed integer value in each element of the four source vectors, and places the four-way interleaved rounded results in the quarter-width destination elements. Each result element is saturated to the quarter-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 source element.

This instruction is unpredicated.

Encoding: SME2

Variants: FEAT_SME2 (ARMv9.3)

313029282726252423222120191817161514131211109876543210
11000001111011100
tsizeimm5NZnopUZd

SQRSHRN <Zd>.<T>, { <Zn1>.<Tb>-<Zn4>.<Tb> }, #<const>

Decoding algorithm

if !IsFeatureImplemented(FEAT_SME2) then EndOfDecode(Decode_UNDEF);
if tsize == '00' then EndOfDecode(Decode_UNDEF);
constant integer esize = 8 << HighestSetBit(tsize);
constant integer n = UInt(Zn:'00');
constant integer d = UInt(Zd);
constant integer shift = (8 * esize) - UInt(tsize:imm5);

Operation

CheckStreamingSVEEnabled();
constant integer VL = CurrentVL;
constant integer elements = VL DIV (4 * esize);
bits(VL) result;

for e = 0 to elements-1
    for i = 0 to 3
        constant bits(VL) operand = Z[n+i, VL];
        constant bits(4 * esize) element = Elem[operand, e, 4 * esize];
        constant integer res = (SInt(element) + (1 << (shift-1))) >> shift;
        Elem[result, 4*e + i, 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>: <Zn1>: Is the name of the first scalable vector register of the source multi-vector group, encoded as "Zn" times 4.
<Tb>: <Zn4>: Is the name of the fourth scalable vector register of the source multi-vector group, encoded as "Zn" times 4 plus 3.
<const>: Is the immediate shift amount, in the range 1 to number of bits per source element, encoded in "tsize:imm5".