Floating-point down convert, rounding to odd (top, predicated)
Convert active double-precision floating-point elements from the source vector to single-precision, rounding to Odd, and place the results in the odd-numbered 32-bit elements of the destination vector, leaving the even-numbered elements unchanged. Inactive elements in the destination vector register remain unmodified.
Rounding to Odd (aka Von Neumann rounding) permits a two-step conversion from double-precision to half-precision without incurring intermediate rounding errors.
| 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 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | Pg | Zn | Zd | ||||||||||
if !HaveSVE2() && !HaveSME() then UNDEFINED; integer esize = 64; integer g = UInt(Pg); integer n = UInt(Zn); integer d = UInt(Zd);
| <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. |
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[], FPRounding_ODD); 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.