VCVTB (BFloat16)

Converts the single-precision value in a single-precision register to BFloat16 format and writes the result into the bottom half of a single precision register, preserving the top 16 bits of the destination register.

Unlike the BFloat16 multiplication instructions, this instruction honors all the control bits in the FPSCR that apply to single-precision arithmetic, including the rounding mode. This instruction can generate a floating-point exception which causes a cumulative exception bit in the FPSCR to be set, or a synchronous exception to be taken, depending on the enable bits in the FPSCR.

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

A1
(FEAT_AA32BF16)

313029282726252423222120191817161514131211109876543210
!= 111111101D110011Vd100101M0Vm
cond

A1

VCVTB{<c>}{<q>}.BF16.F32 <Sd>, <Sm>

if !HaveAArch32BF16Ext() then UNDEFINED; integer d = UInt(Vd:D); integer m = UInt(Vm:M);

T1
(FEAT_AA32BF16)

15141312111098765432101514131211109876543210
111011101D110011Vd100101M0Vm

T1

VCVTB{<c>}{<q>}.BF16.F32 <Sd>, <Sm>

if !HaveAArch32BF16Ext() then UNDEFINED; integer d = UInt(Vd:D); integer 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.

<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); S[d]<15:0> = FPConvertBF(S[m], FPSCR[]);


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.