FMOP4A (non-widening) -- A64

This instruction generates four independent quarter-tile floating-point outer products from the sub-matrices in the half-vectors of the one or two first and second source vectors and accumulates the results to the corresponding elements of a 16-bit, 32-bit, or 64-bit element ZA tile.

In case of the half-precision variant, each of the quarter-tile outer products is generated by multiplying the SVL_H÷2 × 1 sub-matrix of half-precision values held in the half-vectors of the first source vectors by the 1 × SVL_H÷2 sub-matrix of half-precision values held in the half-vectors of the second source vectors. In case of the single-precision variant, each of the quarter-tile outer products is generated by multiplying the SVL_S÷2 × 1 sub-matrix of single-precision values held in the half-vectors of the first source vectors by the 1 × SVL_S÷2 sub-matrix of single-precision values held in the half-vectors of the second source vectors. In case of the double-precision variant, each of the quarter-tile outer products is generated by multiplying the SVL_D÷2 × 1 sub-matrix of double-precision values held in the half-vectors of the first source vectors by the 1 × SVL_D÷2 sub-matrix of double-precision values held in the half-vectors of the second source vectors.

The resulting quarter-tile SVL_H÷2 × SVL_H÷2 half-precision outer products in case of the half-precision variant, SVL_S÷2 × SVL_S÷2 single-precision outer products in case of the single-precision variant, or SVL_D÷2 × SVL_D÷2 double-precision outer products in case of the double-precision variant are destructively added to the destination ZA tile. This is equivalent to performing a single multiply-accumulate to each of the destination tile elements.

Half-precision, single and multiple vectors
(FEAT_SME_MOP4 && FEAT_SME_F16F16)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F16F16) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 16; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Half-precision, single vectors
(FEAT_SME_MOP4 && FEAT_SME_F16F16)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F16F16) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 16; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Half-precision, multiple and single vectors
(FEAT_SME_MOP4 && FEAT_SME_F16F16)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F16F16) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 16; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Half-precision, multiple vectors
(FEAT_SME_MOP4 && FEAT_SME_F16F16)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F16F16) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 16; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Single-precision, single and multiple vectors
(FEAT_SME_MOP4)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 32; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Single-precision, single vectors
(FEAT_SME_MOP4)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 32; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Single-precision, multiple and single vectors
(FEAT_SME_MOP4)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 32; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Single-precision, multiple vectors
(FEAT_SME_MOP4)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 32; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Double-precision, single and multiple vectors
(FEAT_SME_MOP4 && FEAT_SME_F64F64)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F64F64) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 64; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Double-precision, single vectors
(FEAT_SME_MOP4 && FEAT_SME_F64F64)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F64F64) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 64; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 1; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Double-precision, multiple and single vectors
(FEAT_SME_MOP4 && FEAT_SME_F64F64)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F64F64) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 64; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 1; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Double-precision, multiple vectors
(FEAT_SME_MOP4 && FEAT_SME_F64F64)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME_MOP4) || !IsFeatureImplemented(FEAT_SME_F64F64) then EndOfDecode(Decode_UNDEF); end; let esize : integer{} = 64; let n : integer = UInt('0'::Zn::'0'); let m : integer = UInt('1'::Zm::'0'); let nreg : integer{} = 2; let mreg : integer = 2; let da : integer = UInt(ZAda); let sub_op : boolean = FALSE;

Assembler Symbols

<ZAda>	For the "Half-precision, multiple and single vectors", "Half-precision, multiple vectors", "Half-precision, single and multiple vectors", and "Half-precision, single vectors" variants: is the name of the ZA tile ZA0-ZA1, encoded in the "ZAda" field.
	For the "Single-precision, multiple and single vectors", "Single-precision, multiple vectors", "Single-precision, single and multiple vectors", and "Single-precision, single vectors" variants: is the name of the ZA tile ZA0-ZA3, encoded in the "ZAda" field.
	For the "Double-precision, multiple and single vectors", "Double-precision, multiple vectors", "Double-precision, single and multiple vectors", and "Double-precision, single vectors" variants: is the name of the ZA tile ZA0-ZA7, encoded in the "ZAda" field.

<Zn>	Is the name of the first source scalable vector register, registers in the range Z0-Z15, encoded as "Zn" times 2.

<Zm1>

Is the name of the first scalable vector register of the second source multi-vector group, in the range Z16-Z31, encoded as "Zm" times 2 plus 16.

<Zm2>

Is the name of the second scalable vector register of the second source multi-vector group, in the range Z16-Z31, encoded as "Zm" times 2 plus 17.

<Zm>	Is the name of the second source scalable vector register, registers in the range Z16-Z31, encoded as "Zm" times 2 plus 16.

<Zn1>

Is the name of the first scalable vector register of the first source multi-vector group, in the range Z0-Z15, encoded as "Zn" times 2.

<Zn2>

Is the name of the second scalable vector register of the first source multi-vector group, in the range Z0-Z15, encoded as "Zn" times 2 plus 1.

Operation

CheckStreamingSVEAndZAEnabled(); let VL : integer{} = CurrentVL(); let hvsize : integer{} = VL DIV 2; let dim : integer{} = hvsize DIV esize; let tilesize : integer{} = 4*dim*dim*esize; let op3 : bits(tilesize) = ZAtile{}(da, esize); var result : bits(tilesize); for outprod = 0 to 3 do let row_hv : integer = outprod DIVRM 2; let col_hv : integer = outprod MOD 2; let row_base : integer = row_hv * dim; let col_base : integer = col_hv * dim; let op1 : bits(VL) = Z{}(n + (nreg-1)*col_hv); let op2 : bits(VL) = Z{}(m + (mreg-1)*row_hv); for row = 0 to dim-1 do for col = 0 to dim-1 do let row_idx : integer = row_base + row; let col_idx : integer = col_base + col; let tile_idx : integer = row_idx * dim * 2 + col_idx; var elem1 : bits(esize) = op1[row_idx*:(esize)]; let elem2 : bits(esize) = op2[col_idx*:(esize)]; let elem3 : bits(esize) = op3[tile_idx*:(esize)]; if sub_op then elem1 = FPNeg{esize}(elem1, FPCR()); end; result[tile_idx*:(esize)] = FPMulAdd_ZA{esize}(elem3, elem1, elem2, FPCR()); end; end; end; ZAtile{tilesize}(da, esize) = result;

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	0	0	0	0	0	0	1	0	0	0	1	Zm			0	0	0	0	0	0	0	0	Zn			0	0	1	0	0	ZAda
											M											N					S

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	0	0	0	0	0	0	1	0	0	0	0	Zm			0	0	0	0	0	0	0	0	Zn			0	0	1	0	0	ZAda
											M											N					S

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	0	0	0	0	0	0	1	0	0	0	0	Zm			0	0	0	0	0	0	0	1	Zn			0	0	1	0	0	ZAda
											M											N					S

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	0	0	0	0	0	0	1	0	0	0	1	Zm			0	0	0	0	0	0	0	1	Zn			0	0	1	0	0	ZAda
											M											N					S

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	0	0	0	0	0	0	0	0	0	0	1	Zm			0	0	0	0	0	0	0	0	Zn			0	0	0	0	ZAda
											M											N					S