BFCVTNT

Single-precision convert to BFloat16 (top, predicated)

This instruction converts each active floating-point element of the source vector from single-precision to BFloat16 floating-point, and places the results in the odd-numbered 16-bit elements of the destination vector, leaving the even-numbered elements unchanged. Inactive elements in the destination vector register remain unmodified or are set to zero, depending on whether merging or zeroing predication is selected.

ID_AA64ZFR0_EL1.BF16 indicates whether this instruction is implemented.

It has encodings from 2 classes: Merging and Zeroing

Merging
((FEAT_SVE || FEAT_SME) && FEAT_BF16)

313029282726252423222120191817161514131211109876543210
0110010010001010101PgZnZd
opcopc2

Encoding

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

Decode for this encoding

if ((!IsFeatureImplemented(FEAT_SVE) && !IsFeatureImplemented(FEAT_SME)) || !IsFeatureImplemented(FEAT_BF16)) then EndOfDecode(Decode_UNDEF); end; let g : integer = UInt(Pg); let n : integer = UInt(Zn); let d : integer = UInt(Zd); let merging : boolean = TRUE;

Zeroing
(FEAT_SVE2p2 || FEAT_SME2p2)

313029282726252423222120191817161514131211109876543210
0110010010000010101PgZnZd
opcopc2

Encoding

BFCVTNT <Zd>.H, <Pg>/Z, <Zn>.S

Decode for this encoding

if !IsFeatureImplemented(FEAT_SVE2p2) && !IsFeatureImplemented(FEAT_SME2p2) then EndOfDecode(Decode_UNDEF); end; let g : integer = UInt(Pg); let n : integer = UInt(Zn); let d : integer = UInt(Zd); let merging : boolean = FALSE;

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

Operation

CheckSVEEnabled(); let VL : integer{} = CurrentVL(); let PL : integer{} = VL DIV 8; let elements : integer = VL DIV 32; let mask : bits(PL) = P{}(g); let operand : bits(VL) = if AnyActiveElement{PL}(mask, 32) then Z{VL}(n) else Zeros{VL}; var result : bits(VL) = Z{}(d); for e = 0 to elements-1 do if ActivePredicateElement{PL}(mask, e, 32) then let element : bits(32) = operand[e*:32]; result[(2*e+1)*:16] = FPConvertBF(element, FPCR()); elsif !merging then result[(2*e+1)*:16] = Zeros{16}; end; end; Z{VL}(d) = result;


2026-03_rel 2026-03-26 20:48:11

Copyright © 2010-2026 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.