FCVTXNT
Floating-point down convert, rounding to odd (top, predicated)
Convert active double-precision floating-point elements from the source vector to single-precision, rounding to Odd, and place the results in the odd-numbered 32-bit elements of the destination vector, leaving the even-numbered elements unchanged. Inactive elements in the destination vector register remain unmodified.
Rounding to Odd (aka Von Neumann rounding) permits a two-step conversion from double-precision to half-precision without incurring intermediate rounding errors.
Green
True
0
1
1
0
0
1
0
0
0
0
0
0
1
0
1
0
1
0
1
FCVTXNT <Zd>.S, <Pg>/M, <Zn>.D
if !IsFeatureImplemented(FEAT_SVE2) && !IsFeatureImplemented(FEAT_SME) then UNDEFINED;
constant integer esize = 64;
constant integer g = UInt(Pg);
constant integer n = UInt(Zn);
constant integer d = UInt(Zd);
<Zd>
Is the name of the destination scalable vector register, encoded in the "Zd" field.
<Pg>
Is the name of the governing scalable predicate register P0-P7, encoded in the "Pg" field.
<Zn>
Is the name of the source scalable vector register, encoded in the "Zn" field.
CheckSVEEnabled();
constant integer VL = CurrentVL;
constant integer PL = VL DIV 8;
constant integer elements = VL DIV esize;
constant integer halfesize = esize DIV 2;
constant bits(PL) mask = P[g, PL];
constant bits(VL) operand = if AnyActiveElement(mask, esize) then Z[n, VL] else Zeros(VL);
bits(VL) result = Z[d, VL];
for e = 0 to elements-1
if ActivePredicateElement(mask, e, esize) then
constant bits(esize) element = Elem[operand, e, esize];
Elem[result, 2*e + 1, halfesize] = FPConvertSVE(element, FPCR, FPRounding_ODD, halfesize);
Z[d, VL] = result;