PMULL

Multi-vector polynomial multiply long

This instruction performs a polynomial multiplication over [0, 1] of the even-numbered elements of the first source vector by the corresponding elements of the second source vector, and places the results in the overlapping double-width elements of the first destination vector. The same operation is performed with odd-numbered elements of the source vectors, writing to the second destination vector. This instruction is unpredicated.

This instruction is legal when executed in Streaming SVE mode if both FEAT_SSVE_AES and FEAT_SVE_AES2 are implemented.

SVE2
(FEAT_SVE_AES2)

313029282726252423222120191817161514131211109876543210
01000101001Zm111110ZnZd0
size

Encoding

PMULL { <Zd1>.Q-<Zd2>.Q }, <Zn>.D, <Zm>.D

Decode for this encoding

if !IsFeatureImplemented(FEAT_SVE_AES2) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 128; let n : integer = UInt(Zn); let m : integer = UInt(Zm); let d : integer = UInt(Zd::'0');

Assembler Symbols

<Zd1>

Is the name of the first scalable vector register of the destination multi-vector group, encoded as "Zd" times 2.

<Zd2>

Is the name of the second scalable vector register of the destination multi-vector group, encoded as "Zd" times 2 plus 1.

<Zn>

Is the name of the first source scalable vector register, encoded in the "Zn" field.

<Zm>

Is the name of the second source scalable vector register, encoded in the "Zm" field.

Operation

if IsFeatureImplemented(FEAT_SSVE_AES) then CheckSVEEnabled(); else CheckNonStreamingSVEEnabled(); end; let VL : integer{} = CurrentVL(); let elements : integer = VL DIV esize; let operand1 : bits(VL) = Z{}(n); let operand2 : bits(VL) = Z{}(m); var result_lo : bits(VL); var result_hi : bits(VL); for e = 0 to elements-1 do let element1_lo : bits(esize DIV 2) = operand1[(2*e + 0)*:(esize DIV 2)]; let element2_lo : bits(esize DIV 2) = operand2[(2*e + 0)*:(esize DIV 2)]; let element1_hi : bits(esize DIV 2) = operand1[(2*e + 1)*:(esize DIV 2)]; let element2_hi : bits(esize DIV 2) = operand2[(2*e + 1)*:(esize DIV 2)]; result_lo[e*:esize] = PolynomialMult{esize DIV 2, esize DIV 2}(element1_lo, element2_lo); result_hi[e*:esize] = PolynomialMult{esize DIV 2, esize DIV 2}(element1_hi, element2_hi); end; Z{VL}(d + 0) = result_lo; Z{VL}(d + 1) = result_hi;


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.