Floating-point convert to lower precision narrow, rounding to odd (vector)
This instruction reads each double-precision element in the source SIMD&FP register, converts each value to single-precision format using the Round to Odd rounding mode, and writes the results to the destination SIMD&FP register.
This instruction uses the Round to Odd rounding mode, which is not defined by the IEEE 754-2008 standard. This rounding mode ensures that if the result of the conversion is inexact, the least significant bit of the mantissa is forced to 1. This rounding mode enables a floating-point value to be converted to a lower precision format via an intermediate precision format while avoiding double rounding errors. For example, a 64-bit floating-point value can be converted to a correctly rounded 16-bit floating-point value by first using this instruction to produce a 32-bit value and then using another instruction with the wanted rounding mode to convert the 32-bit value to the final 16-bit floating-point value.
The Vector variant of FCVTXN writes the half-width results to the lower half of the destination register and clears the upper half to 0. FCVTXN2 writes the half-width results to the upper half of the destination register without affecting the other bits of the register.
This instruction can generate a floating-point exception. Depending on the settings in FPCR, the exception results in either a flag being set in FPSR or a synchronous exception being generated. For more information, see Floating-point exceptions and exception traps.
Depending on the settings in the CPACR_EL1, CPTR_EL2, and CPTR_EL3 registers, and the current Security state and Exception level, an attempt to execute the instruction might be trapped.
It has encodings from 2 classes: Scalar and Vector
| 31 | 30 | 29 | 28 | 27 | 26 | 25 | 24 | 23 | 22 | 21 | 20 | 19 | 18 | 17 | 16 | 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
| 0 | 1 | 1 | 1 | 1 | 1 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | Rn | Rd | ||||||||
| U | size | opcode | |||||||||||||||||||||||||||||
if !IsFeatureImplemented(FEAT_AdvSIMD) then EndOfDecode(Decode_UNDEF); constant integer d = UInt(Rd); constant integer n = UInt(Rn); constant integer esize = 32; constant integer datasize = esize; constant integer elements = 1; constant integer part = 0;
| 31 | 30 | 29 | 28 | 27 | 26 | 25 | 24 | 23 | 22 | 21 | 20 | 19 | 18 | 17 | 16 | 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
| 0 | Q | 1 | 0 | 1 | 1 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | Rn | Rd | ||||||||
| U | size | opcode | |||||||||||||||||||||||||||||
if !IsFeatureImplemented(FEAT_AdvSIMD) then EndOfDecode(Decode_UNDEF); constant integer d = UInt(Rd); constant integer n = UInt(Rn); constant integer esize = 32; constant integer datasize = 64; constant integer elements = 2; constant integer part = UInt(Q);
| <d> |
Is the number of the SIMD&FP destination register, encoded in the "Rd" field. |
| <n> |
Is the number of the SIMD&FP source register, encoded in the "Rn" field. |
| <Vd> |
Is the name of the SIMD&FP destination register, encoded in the "Rd" field. |
| <Tb> |
Is an arrangement specifier,
encoded in
|
| <Vn> |
Is the name of the SIMD&FP source register, encoded in the "Rn" field. |
AArch64.CheckFPAdvSIMDEnabled(); constant bits(2*datasize) operand = V[n, 2*datasize]; constant boolean merge = elements == 1 && IsMerging(FPCR); bits(128) result = if merge then V[d, 128] else Zeros(128); for e = 0 to elements-1 Elem[result, e, esize] = FPConvert(Elem[operand, e, 2*esize], FPCR, FPRounding_ODD, esize); if merge then V[d, 128] = result; else Vpart[d, part, datasize] = Elem[result, 0, datasize];
Version 2025.09 — Copyright © 2010-2025 Arm Limited or its affiliates.
This site is provided as a community resource and is NOT affiliated with nor endorsed by Arm Limited.