FRSQRTS

Floating-point reciprocal square root step (unpredicated)

Multiply corresponding floating-point elements of the first and second source vectors, subtract the products from 3.0 and divide the results by 2.0 without any intermediate rounding and place the results in the corresponding elements of the destination vector. This instruction is unpredicated.

This instruction can be used to perform a single Newton-Raphson iteration for calculating the reciprocal square root of a vector of floating-point values.

Encoding: SVE

Variants: FEAT_SVE || FEAT_SME (FEAT_SVE || FEAT_SME)

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

FRSQRTS <Zd>.<T>, <Zn>.<T>, <Zm>.<T>

Decoding algorithm

if !IsFeatureImplemented(FEAT_SVE) && !IsFeatureImplemented(FEAT_SME) then
    EndOfDecode(Decode_UNDEF);
if size == '00' then EndOfDecode(Decode_UNDEF);
constant integer esize = 8 << UInt(size);
constant integer n = UInt(Zn);
constant integer m = UInt(Zm);
constant integer d = UInt(Zd);

Operation

CheckSVEEnabled();
constant integer VL = CurrentVL;
constant integer elements = VL DIV esize;
constant bits(VL) operand1 = Z[n, VL];
constant bits(VL) operand2 = Z[m, VL];
bits(VL) result;

for e = 0 to elements-1
    constant bits(esize) element1 = Elem[operand1, e, esize];
    constant bits(esize) element2 = Elem[operand2, e, esize];
    Elem[result, e, esize] = FPRSqrtStepFused(element1, element2, FPCR);

Z[d, VL] = result;

Explanations

<Zd>: Is the name of the destination scalable vector register, encoded in the "Zd" field.
<T>: <Zn>: Is the name of the first source scalable vector register, encoded in the "Zn" field.
<Zm>: Is the name of the second source scalable vector register, encoded in the "Zm" field.