FCVTNT

Floating-point down convert and narrow (top, predicated)

Convert active floating-point elements from the source vector to the next lower precision, and place the results in the odd-numbered half-width elements of the destination vector, leaving the even-numbered elements unchanged. Inactive elements in the destination vector register remain unmodified.

It has encodings from 2 classes: Single-precision to half-precision and Double-precision to single-precision

Single-precision to half-precision

313029282726252423222120191817161514131211109876543210
0110010010001000101PgZnZd

FCVTNT <Zd>.H, <Pg>/M, <Zn>.S

if !HaveSVE2() && !HaveSME() then UNDEFINED; integer esize = 32; integer g = UInt(Pg); integer n = UInt(Zn); integer d = UInt(Zd);

Double-precision to single-precision

313029282726252423222120191817161514131211109876543210
0110010011001010101PgZnZd

FCVTNT <Zd>.S, <Pg>/M, <Zn>.D

if !HaveSVE2() && !HaveSME() then UNDEFINED; integer esize = 64; integer g = UInt(Pg); integer n = UInt(Zn); integer d = UInt(Zd);

Assembler Symbols

<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 first source scalable vector register, encoded in the "Zn" field.

Operation

CheckSVEEnabled(); integer elements = VL DIV esize; bits(PL) mask = P[g]; bits(VL) operand = if AnyActiveElement(mask, esize) then Z[n] else Zeros(); bits(VL) result = Z[d]; for e = 0 to elements-1 if ElemP[mask, e, esize] == '1' then bits(esize) element = Elem[operand, e, esize]; Elem[result, 2*e + 1, esize DIV 2] = FPConvertSVE(element, FPCR[]); Z[d] = result;


Internal version only: isa v33.11seprel, AdvSIMD v29.05, pseudocode v2021-09_rel, sve v2021-09_rc3d ; Build timestamp: 2021-10-06T11:41

Copyright © 2010-2021 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.