SDOT (2-way, multiple vectors)

Multi-vector signed integer dot-product

The signed integer dot product instruction computes the dot product of two signed 16-bit integer values held in each 32-bit element of the two or four first source vectors and two signed 16-bit integer values in the corresponding 32-bit element of the two or four second source vectors. The widened dot product result is destructively added to the corresponding 32-bit element of the ZA single-vector groups.

The single-vector group within each half of or each quarter of the ZA array is selected by the sum of the vector select register and offset, modulo half or quarter the number of ZA array vectors.

The vector group symbol, VGx2 or VGx4, indicates that the ZA operand consists of two or four ZA single-vector groups respectively. The vector group symbol is preferred for disassembly, but optional in assembler source code.

This instruction is unpredicated.

It has encodings from 2 classes: Two ZA single-vectors and Four ZA single-vectors

Two ZA single-vectors
(FEAT_SME2)

313029282726252423222120191817161514131211109876543210
11000001111Zm00Rv101Zn001off3
U

SDOT ZA.S[<Wv>, <offs>{, VGx2}], { <Zn1>.H-<Zn2>.H }, { <Zm1>.H-<Zm2>.H }

if !IsFeatureImplemented(FEAT_SME2) then UNDEFINED; constant integer v = UInt('010':Rv); constant integer esize = 32; constant integer n = UInt(Zn:'0'); constant integer m = UInt(Zm:'0'); constant integer offset = UInt(off3); constant integer nreg = 2;

Four ZA single-vectors
(FEAT_SME2)

313029282726252423222120191817161514131211109876543210
11000001111Zm010Rv101Zn0001off3
U

SDOT ZA.S[<Wv>, <offs>{, VGx4}], { <Zn1>.H-<Zn4>.H }, { <Zm1>.H-<Zm4>.H }

if !IsFeatureImplemented(FEAT_SME2) then UNDEFINED; constant integer v = UInt('010':Rv); constant integer esize = 32; constant integer n = UInt(Zn:'00'); constant integer m = UInt(Zm:'00'); constant integer offset = UInt(off3); constant integer nreg = 4;

Assembler Symbols

<Wv>

Is the 32-bit name of the vector select register W8-W11, encoded in the "Rv" field.

<offs>

Is the vector select offset, in the range 0 to 7, encoded in the "off3" field.

<Zn1>

For the two ZA single-vectors variant: is the name of the first scalable vector register of the first source multi-vector group, encoded as "Zn" times 2.

For the four ZA single-vectors variant: is the name of the first scalable vector register of the first source multi-vector group, encoded as "Zn" times 4.

<Zn4>

Is the name of the fourth scalable vector register of the first source multi-vector group, encoded as "Zn" times 4 plus 3.

<Zn2>

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

<Zm1>

For the two ZA single-vectors variant: is the name of the first scalable vector register of the second source multi-vector group, encoded as "Zm" times 2.

For the four ZA single-vectors variant: is the name of the first scalable vector register of the second source multi-vector group, encoded as "Zm" times 4.

<Zm4>

Is the name of the fourth scalable vector register of the second source multi-vector group, encoded as "Zm" times 4 plus 3.

<Zm2>

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

Operation

CheckStreamingSVEAndZAEnabled(); constant integer VL = CurrentVL; constant integer elements = VL DIV esize; constant integer vectors = VL DIV 8; constant integer vstride = vectors DIV nreg; constant bits(32) vbase = X[v, 32]; integer vec = (UInt(vbase) + offset) MOD vstride; bits(VL) result; for r = 0 to nreg-1 constant bits(VL) operand1 = Z[n+r, VL]; constant bits(VL) operand2 = Z[m+r, VL]; constant bits(VL) operand3 = ZAvector[vec, VL]; for e = 0 to elements-1 bits(esize) sum = Elem[operand3, e, esize]; for i = 0 to 1 constant integer element1 = SInt(Elem[operand1, 2 * e + i, esize DIV 2]); constant integer element2 = SInt(Elem[operand2, 2 * e + i, esize DIV 2]); sum = sum + element1 * element2; Elem[result, e, esize] = sum; ZAvector[vec, VL] = result; vec = vec + vstride;


Internal version only: aarchmrs v2024-03_relA, pseudocode v2024-03_rel, sve v2024-03_rel ; Build timestamp: 2024-03-26T09:45

Copyright © 2010-2024 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.