Multi-vector signed rounding shift left by vector
This instruction shifts the signed elements of the two or four first source vectors by the corresponding elements of the second source vector, and destructively places the rounded results in the corresponding elements of the two or four first source vectors. A positive shift amount performs a left shift, otherwise a right shift by the negated shift amount is performed.
This instruction is unpredicated.
It has encodings from 2 classes: Two registers and Four registers
| 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 |
| 1 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | size | 1 | 0 | Zm | 1 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | Zdn | 0 | |||||||
| opc | U | ||||||||||||||||||||||||||||||
if !IsFeatureImplemented(FEAT_SME2) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 8 << UInt(size); let dn : integer = UInt(Zdn::'0'); let m : integer = UInt('0'::Zm); let nreg : integer{} = 2;
| 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 |
| 1 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | size | 1 | 0 | Zm | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | Zdn | 0 | 0 | ||||||
| opc | U | ||||||||||||||||||||||||||||||
if !IsFeatureImplemented(FEAT_SME2) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 8 << UInt(size); let dn : integer = UInt(Zdn::'00'); let m : integer = UInt('0'::Zm); let nreg : integer{} = 4;
| <T> |
Is the size specifier,
encoded in
|
| <Zdn2> |
Is the name of the second scalable vector register of the destination and first source multi-vector group, encoded as "Zdn" times 2 plus 1. |
| <Zm> |
Is the name of the second source scalable vector register Z0-Z15, encoded in the "Zm" field. |
| <Zdn4> |
Is the name of the fourth scalable vector register of the destination and first source multi-vector group, encoded as "Zdn" times 4 plus 3. |
CheckStreamingSVEEnabled(); let VL : integer{} = CurrentVL(); let elements : integer = VL DIV esize; var results : array [[4]] of bits(VL); for r = 0 to nreg-1 do let operand1 : bits(VL) = Z{}(dn+r); let operand2 : bits(VL) = Z{}(m); for e = 0 to elements-1 do let element : integer = SInt(operand1[e*:esize]); var shift : integer = ShiftSat(SInt(operand2[e*:esize]), esize); var res : integer; if shift >= 0 then res = element << shift; else shift = -shift; res = (element + (1 << (shift - 1))) >> shift; end; results[[r]][e*:esize] = res[esize-1:0]; end; end; for r = 0 to nreg-1 do Z{VL}(dn+r) = results[[r]]; end;
2026-03_rel 2026-03-26 20:48:11
Copyright © 2010-2026 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.