VCVTB

Convert to or from a half-precision value in the bottom half of a single-precision register does one of the following:

Depending on settings in the CPACR, NSACR, HCPTR, and FPEXC registers, and the Security state and PE mode in which the instruction is executed, an attempt to execute the instruction might be undefined, or trapped to Hyp mode. For more information see Enabling Advanced SIMD and floating-point support.

It has encodings from the following instruction sets: A32 ( A1 ) and T32 ( T1 ) .

A1

313029282726252423222120191817161514131211109876543210
!= 111111101D11001opVd101sz01M0Vm
condT

Half-precision to single-precision (op == 0 && sz == 0)

VCVTB{<c>}{<q>}.F32.F16 <Sd>, <Sm>

Half-precision to double-precision (op == 0 && sz == 1)

VCVTB{<c>}{<q>}.F64.F16 <Dd>, <Sm>

Single-precision to half-precision (op == 1 && sz == 0)

VCVTB{<c>}{<q>}.F16.F32 <Sd>, <Sm>

Double-precision to half-precision (op == 1 && sz == 1)

VCVTB{<c>}{<q>}.F16.F64 <Sd>, <Dm>

uses_double = (sz == '1'); convert_from_half = (op == '0'); lowbit = (if T == '1' then 16 else 0); integer d; integer m; if uses_double then if convert_from_half then d = UInt(D:Vd); m = UInt(Vm:M); else d = UInt(Vd:D); m = UInt(M:Vm); else d = UInt(Vd:D); m = UInt(Vm:M);

T1

15141312111098765432101514131211109876543210
111011101D11001opVd101sz01M0Vm
T

Half-precision to single-precision (op == 0 && sz == 0)

VCVTB{<c>}{<q>}.F32.F16 <Sd>, <Sm>

Half-precision to double-precision (op == 0 && sz == 1)

VCVTB{<c>}{<q>}.F64.F16 <Dd>, <Sm>

Single-precision to half-precision (op == 1 && sz == 0)

VCVTB{<c>}{<q>}.F16.F32 <Sd>, <Sm>

Double-precision to half-precision (op == 1 && sz == 1)

VCVTB{<c>}{<q>}.F16.F64 <Sd>, <Dm>

uses_double = (sz == '1'); convert_from_half = (op == '0'); lowbit = (if T == '1' then 16 else 0); integer d; integer m; if uses_double then if convert_from_half then d = UInt(D:Vd); m = UInt(Vm:M); else d = UInt(Vd:D); m = UInt(M:Vm); else d = UInt(Vd:D); m = UInt(Vm:M);

Assembler Symbols

<c>

See Standard assembler syntax fields.

<q>

See Standard assembler syntax fields.

<Sd>

Is the 32-bit name of the SIMD&FP destination register, encoded in the "Vd:D" field.

<Dm>

Is the 64-bit name of the SIMD&FP source register, encoded in the "M:Vm" field.

<Dd>

Is the 64-bit name of the SIMD&FP destination register, encoded in the "D:Vd" field.

<Sm>

Is the 32-bit name of the SIMD&FP source register, encoded in the "Vm:M" field.

Operation

if ConditionPassed() then EncodingSpecificOperations(); CheckVFPEnabled(TRUE); bits(16) hp; if convert_from_half then hp = S[m]<lowbit+15:lowbit>; if uses_double then D[d] = FPConvert(hp, FPSCR[], 64); else S[d] = FPConvert(hp, FPSCR[], 32); else if uses_double then hp = FPConvert(D[m], FPSCR[], 16); else hp = FPConvert(S[m], FPSCR[], 16); S[d]<lowbit+15:lowbit> = hp;


Internal version only: isa v01_31, pseudocode v2023-06_rel, sve v2023-06_rel ; Build timestamp: 2023-07-04T18:06

Copyright © 2010-2023 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.