This is a collection of Intel®’ IA32® Software Developer's Manuals (URL of the day) and AMD' AMD64 Architecture Programmer's Manual together with the related specifications, application notes, white papers, and change logs. The collection aims to keep all available revisions. It was originally created by Michal Necasek, see OS/2 Museum.

If you have a public document, related to the IA32® specifications and missing from the collection, please mail it to me. The content of this URL and all sub-ULRs is available for convenient bulk download by rsync x86docs password "" (empty).

SMMLA, SMMLAR -- AArch32

SMMLA, SMMLAR

Signed Most Significant Word Multiply Accumulate multiplies two signed 32-bit values, extracts the most significant 32 bits of the result, and adds an accumulate value.

Optionally, the instruction can specify that the result is rounded instead of being truncated. In this case, the constant 0x80000000 is added to the product before the high word is extracted.

It has encodings from the following instruction sets: A32 ( A1 ) and T32 ( T1 ) .

A1

313029282726252423222120191817161514131211109876543210
!= 111101110101Rd!= 1111Rm00R1Rn
condop1Ra

Encoding for the SMMLA variant

Applies when (R == 0)

SMMLA{<c>}{<q>} <Rd>, <Rn>, <Rm>, <Ra>

Encoding for the SMMLAR variant

Applies when (R == 1)

SMMLAR{<c>}{<q>} <Rd>, <Rn>, <Rm>, <Ra>

Decode for all variants of this encoding

if Ra == '1111' then See("SMMUL"); end; let d : integer = UInt(Rd); let n : integer = UInt(Rn); let m : integer = UInt(Rm); let a : integer = UInt(Ra); let round : boolean = (R == '1'); if d == 15 || n == 15 || m == 15 then UnpredictableProcedure(); end;

T1

15141312111098765432101514131211109876543210
111110110101Rn!= 1111Rd000RRm
op1Ra

Encoding for the SMMLA variant

Applies when (R == 0)

SMMLA{<c>}{<q>} <Rd>, <Rn>, <Rm>, <Ra>

Encoding for the SMMLAR variant

Applies when (R == 1)

SMMLAR{<c>}{<q>} <Rd>, <Rn>, <Rm>, <Ra>

Decode for all variants of this encoding

if Ra == '1111' then See("SMMUL"); end; let d : integer = UInt(Rd); let n : integer = UInt(Rn); let m : integer = UInt(Rm); let a : integer = UInt(Ra); let round : boolean = (R == '1'); // Armv8-A removes UNPREDICTABLE for R13 if d == 15 || n == 15 || m == 15 then UnpredictableProcedure(); end;

For more information about the CONSTRAINED UNPREDICTABLE behavior of this instruction, see Architectural Constraints on UNPREDICTABLE behaviors.

Assembler Symbols

<c>

See Standard assembler syntax fields.

<q>

See Standard assembler syntax fields.

<Rd>

Is the general-purpose destination register, encoded in the "Rd" field.

<Rn>

Is the first general-purpose source register holding the multiplicand, encoded in the "Rn" field.

<Rm>

Is the second general-purpose source register holding the multiplier, encoded in the "Rm" field.

<Ra>

Is the third general-purpose source register holding the addend, encoded in the "Ra" field.

Operation

if ConditionPassed() then EncodingSpecificOperations(); var result : integer = (SInt(R(a)) << 32) + SInt(R(n)) * SInt(R(m)); if round then result = result + 0x80000000; end; R(d) = result[63:32]; end;

Operational information

This instruction is a data-independent-time instruction as described in About the DIT bit.


2025-09_rel_asl1 2026-03-12 12:57:38

Copyright © 2010-2025 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.