st1w
SVE Scatter Store Words (Vector Index)
ST1W { <Zt>.S }, <Pg>, [<Xn|SP>, <Zm>.S, SXTW #<shift>]
Stores words to non-contiguous addresses.
Details
Performs a scatter-gather store of 32-bit words from Zt to non-contiguous memory addresses computed by adding scaled vector indices to a base address. Elements are stored only where the corresponding predicate bit in Pg is set. Does not modify condition flags. Execution restricted to AArch64 with SVE extension; may generate memory-access exceptions.
Pseudocode Operation
for i = 0 to VL/32-1
if Pg[i] == 1
addr ← Xn + (Zm[i] << shift)
[addr, 32] ← Zt[i, 32]
Example
ST1W p0/m, [x1, z2.s.S, SXTW #LSL]
Encoding
Binary Layout
1110010
1
0
11
Zm
1
xs
0
Pg
Rn
Zt
Operands
-
Zt
Transfer scalable vector register (SVE load/store) -
Pg
Mask -
Xn
First source / base 64-bit integer register -
Zm
Indices
Reference (Arm A64 ISA)
Instruction Forms
| Encoding | Instruction | ISA | Bit pattern | ||
|---|---|---|---|---|---|
| 0xA0604000 | ST1W { <Zt1>.S-<Zt2>.S }, <PNg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 101000000110 | imm4 | 0 | 1 | 0 | PNg | Rn | Zt | 0 | ||
| 0xA060C000 | ST1W { <Zt1>.S-<Zt4>.S }, <PNg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 101000000110 | imm4 | 1 | 1 | 0 | PNg | Rn | Zt | 0 | 0 | ||
| 0xA0204000 | ST1W { <Zt1>.S-<Zt2>.S }, <PNg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 10100000001 | Rm | 0 | 1 | 0 | PNg | Rn | Zt | 0 | ||
| 0xA020C000 | ST1W { <Zt1>.S-<Zt4>.S }, <PNg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 10100000001 | Rm | 1 | 1 | 0 | PNg | Rn | Zt | 0 | 0 | ||
| 0xA1604000 | ST1W { <Zt1>.S, <Zt2>.S }, <PNg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 101000010110 | imm4 | 0 | 1 | 0 | PNg | Rn | T | 0 | Zt | ||
| 0xA160C000 | ST1W { <Zt1>.S, <Zt2>.S, <Zt3>.S, <Zt4>.S }, <PNg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 101000010110 | imm4 | 1 | 1 | 0 | PNg | Rn | T | 0 | 0 | Zt | ||
| 0xA1204000 | ST1W { <Zt1>.S, <Zt2>.S }, <PNg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 10100001001 | Rm | 0 | 1 | 0 | PNg | Rn | T | 0 | Zt | ||
| 0xA120C000 | ST1W { <Zt1>.S, <Zt2>.S, <Zt3>.S, <Zt4>.S }, <PNg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 10100001001 | Rm | 1 | 1 | 0 | PNg | Rn | T | 0 | 0 | Zt | ||
| 0xE560A000 | ST1W { <Zt>.S }, <Pg>, [<Zn>.S{, #<imm>}] | A64 | 1110010 | 1 | 0 | 11 | imm5 | 101 | Pg | Zn | Zt | ||
| 0xE540A000 | ST1W { <Zt>.D }, <Pg>, [<Zn>.D{, #<imm>}] | A64 | 1110010 | 1 | 0 | 10 | imm5 | 101 | Pg | Zn | Zt | ||
| 0xE540E000 | ST1W { <Zt>.<T> }, <Pg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 1110010 | 1 | 0 | 1 | sz | 0 | imm4 | 111 | Pg | Rn | Zt | ||
| 0xE500E000 | ST1W { <Zt>.Q }, <Pg>, [<Xn|SP>{, #<imm>, MUL VL}] | A64 | 1110010 | 1 | 0 | 00 | 0 | imm4 | 111 | Pg | Rn | Zt | ||
| 0xE5404000 | ST1W { <Zt>.<T> }, <Pg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 1110010 | 10 | 1 | sz | Rm | 010 | Pg | Rn | Zt | ||
| 0xE5004000 | ST1W { <Zt>.Q }, <Pg>, [<Xn|SP>, <Xm>, LSL #2] | A64 | 1110010 | 10 | 0 | 0 | Rm | 010 | Pg | Rn | Zt |
Description
Scatter store of words from the active elements of a vector register to the memory addresses generated by a 64-bit scalar base plus vector index. The index values are optionally first sign or zero-extended from 32 to 64 bits and then optionally multiplied by 4. Inactive elements are not written to memory.
This instruction is illegal when executed in Streaming SVE mode, unless FEAT_SME_FA64 is implemented and enabled.
Operation
CheckNonStreamingSVEEnabled();
constant integer VL = CurrentVL;
constant integer PL = VL DIV 8;
constant integer elements = VL DIV esize;
bits(64) base;
bits(PL) mask = P[g, PL];
bits(VL) offset;
bits(VL) src;
constant integer mbytes = msize DIV 8;
boolean contiguous = FALSE;
boolean nontemporal = FALSE;
boolean tagchecked = TRUE;
AccessDescriptor accdesc = CreateAccDescSVE(MemOp_STORE, nontemporal, contiguous, tagchecked);
if !AnyActiveElement(mask, esize) then
if n == 31 && ConstrainUnpredictableBool(Unpredictable_CHECKSPNONEACTIVE) then
CheckSPAlignment();
else
if n == 31 then CheckSPAlignment();
base = if n == 31 then SP[] else X[n, 64];
offset = Z[m, VL];
src = Z[t, VL];
for e = 0 to elements-1
if ActivePredicateElement(mask, e, esize) then
integer off = Int(Elem[offset, e, esize]<offs_size-1:0>, offs_unsigned);
bits(64) addr = GenerateAddress(base, off << scale, accdesc);
Mem[addr, mbytes, accdesc] = Elem[src, e, esize]<msize-1:0>;