You copied the Doc URL to your clipboard.

PRFB (scalar plus immediate)

Contiguous prefetch bytes (immediate index).

Contiguous prefetch of byte elements from the memory address generated by a 64-bit scalar base and immediate index in the range -32 to 31 which is multiplied by the vector's in-memory size, irrespective of predication, and added to the base address.

The predicate may be used to suppress prefetches from unwanted addresses.

313029282726252423222120191817161514131211109876543210
1000010111imm6000PgRn0prfop
msz<1>msz<0>

SVE

PRFB <prfop>, <Pg>, [<Xn|SP>{, #<imm>, MUL VL}]

if !HaveSVE() then UNDEFINED;
integer esize = 8;
integer g = UInt(Pg);
integer n = UInt(Rn);
integer level = UInt(prfop<2:1>);
boolean stream = (prfop<0> == '1');
pref_hint = if prfop<3> == '0' then Prefetch_READ else Prefetch_WRITE;
integer scale = 0;
integer offset = SInt(imm6);

Assembler Symbols

<prfop> Is the prefetch operation specifier, encoded in prfop:
prfop <prfop>
0000 PLDL1KEEP
0001 PLDL1STRM
0010 PLDL2KEEP
0011 PLDL2STRM
0100 PLDL3KEEP
0101 PLDL3STRM
x11x #uimm4
1000 PSTL1KEEP
1001 PSTL1STRM
1010 PSTL2KEEP
1011 PSTL2STRM
1100 PSTL3KEEP
1101 PSTL3STRM
<Pg>

Is the name of the governing scalable predicate register P0-P7, encoded in the "Pg" field.

<Xn|SP>

Is the 64-bit name of the general-purpose base register or stack pointer, encoded in the "Rn" field.

<imm>

Is the optional signed immediate vector offset, in the range -32 to 31, defaulting to 0, encoded in the "imm6" field.

Operation

CheckSVEEnabled();
integer elements = VL DIV esize;
bits(PL) mask = P[g];
bits(64) base;
bits(64) addr;

if n == 31 then
    base = SP[];
else
    base = X[n];

addr = base + ((offset * elements) << scale);
for e = 0 to elements-1
    if ElemP[mask, e, esize] == '1' then
        Hint_Prefetch(addr, pref_hint, level, stream);
    addr = addr + (1 << scale);
Previous Next