Splice two vectors under predicate control
This instruction selects a region from the first source vector and copies it to the lowest-numbered elements of the result. Any remaining elements of the result are set to a copy of the lowest-numbered elements from the second source vector. The region is selected using the first and last true elements in the vector select predicate register. The result is destructively placed in the destination and first source vector, or constructively placed in the destination vector.
It has encodings from 2 classes: Constructive and Destructive
| 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 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | size | 1 | 0 | 1 | 1 | 0 | 1 | 1 | 0 | 0 | Pv | Zn | Zd | |||||||||||
if !IsFeatureImplemented(FEAT_SVE2) && !IsFeatureImplemented(FEAT_SME) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 8 << UInt(size); let v : integer = UInt(Pv); let dst : integer = UInt(Zd); let s1 : integer = UInt(Zn); let s2 : integer = (s1 + 1) MOD 32;
| 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 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | size | 1 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | Pv | Zm | Zdn | |||||||||||
if !IsFeatureImplemented(FEAT_SVE) && !IsFeatureImplemented(FEAT_SME) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 8 << UInt(size); let v : integer = UInt(Pv); let dst : integer = UInt(Zdn); let s1 : integer = dst; let s2 : integer = UInt(Zm);
| <Zd> |
Is the name of the destination scalable vector register, encoded in the "Zd" field. |
| <T> |
Is the size specifier,
encoded in
|
| <Pv> |
Is the name of the vector select predicate register P0-P7, encoded in the "Pv" field. |
| <Zn1> |
Is the name of the first scalable vector register of the source multi-vector group, encoded in the "Zn" field. |
| <Zn2> |
Is the name of the second scalable vector register of the source multi-vector group, encoded in the "Zn" field. |
| <Zdn> |
Is the name of the first source and destination scalable vector register, encoded in the "Zdn" field. |
| <Zm> |
Is the name of the second source scalable vector register, encoded in the "Zm" field. |
CheckSVEEnabled(); let VL : integer{} = CurrentVL(); let PL : integer{} = VL DIV 8; let elements : integer = VL DIV esize; let mask : bits(PL) = P{}(v); let operand1 : bits(VL) = if AnyActiveElement{PL}(mask, esize) then Z{VL}(s1) else Zeros{VL}; let operand2 : bits(VL) = Z{}(s2); var result : bits(VL); var x : integer = 0; var active : boolean = FALSE; let lastnum : integer = LastActiveElement{PL}(mask, esize); if lastnum >= 0 then for e = 0 to lastnum do active = active || ActivePredicateElement{PL}(mask, e, esize); if active then result[x*:esize] = operand1[e*:esize]; x = x + 1; end; end; end; let nelements : integer = (elements - x) - 1; for e = 0 to nelements do result[x*:esize] = operand2[e*:esize]; x = x + 1; end; Z{VL}(dst) = result;
For the "Destructive" variant:
The destructive variant of this instruction might be immediately preceded in program order by a MOVPRFX instruction. The MOVPRFX must conform to all of the following requirements, otherwise the behavior of the MOVPRFX and the destructive variant of this instruction is CONSTRAINED UNPREDICTABLE:
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.