BFMUL (multiple vectors) -- A64

This instruction multiplies all the BFloat16 elements of the two or four first source vectors by the corresponding BFloat16 elements of the two or four second source vectors, and places the results in the corresponding elements of the two or four destination vectors.

This instruction follows SME2 non-widening BFloat16 numerical behaviors corresponding to instructions that place their results in two or four SVE Z vectors.

Two registers
(FEAT_SME2 && FEAT_SVE_BFSCALE)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME2) || !IsFeatureImplemented(FEAT_SVE_BFSCALE) then EndOfDecode(Decode_UNDEF); end; let d : integer = UInt(Zd::'0'); let n : integer = UInt(Zn::'0'); let m : integer = UInt(Zm::'0'); let nreg : integer{} = 2;

Four registers
(FEAT_SME2 && FEAT_SVE_BFSCALE)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME2) || !IsFeatureImplemented(FEAT_SVE_BFSCALE) then EndOfDecode(Decode_UNDEF); end; let d : integer = UInt(Zd::'00'); let n : integer = UInt(Zn::'00'); let m : integer = UInt(Zm::'00'); let nreg : integer{} = 4;

Assembler Symbols

<Zd1>	For the "Two registers" variant: is the name of the first scalable vector register of the destination multi-vector group, encoded as "Zd" times 2.
	For the "Four registers" variant: is the name of the first scalable vector register of the destination multi-vector group, encoded as "Zd" times 4.

<Zd2>

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

<Zn1>	For the "Two registers" 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 registers" variant: is the name of the first scalable vector register of the first source multi-vector group, encoded as "Zn" times 4.

<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 registers" 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 registers" variant: is the name of the first scalable vector register of the second source multi-vector group, encoded as "Zm" times 4.

<Zm2>

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

<Zd4>

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

<Zn4>

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

<Zm4>

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

Operation

CheckStreamingSVEEnabled(); let VL : integer{} = CurrentVL(); let elements : integer = VL DIV 16; var results : array [[4]] of bits(VL); for r = 0 to nreg-1 do let operand1 : bits(VL) = Z{}(n+r); let operand2 : bits(VL) = Z{}(m+r); for e = 0 to elements-1 do let element1 : bits(16) = operand1[e*:(16)]; let element2 : bits(16) = operand2[e*:(16)]; results[[r]][e*:(16)] = BFMul(element1, element2, FPCR()); end; end; for r = 0 to nreg-1 do Z{VL}(d+r) = results[[r]]; end;

2026-03_rel 2026-03-26 20:48:11

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	0	0	1	Zm				0	1	1	1	0	0	1	Zn				0	Zd				0
								size

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	0	0	1	Zm			0	1	1	1	1	0	0	1	Zn			0	0	Zd			0	0
								size