8-bit floating-point indexed dot product to single-precision
This instruction computes the fused sum-of-products of a group of four 8-bit floating-point values held in each 32-bit element of the first source vector and a group of four 8-bit floating-point values in an indexed 32-bit element of the second source vector. The single-precision sum-of-products are scaled by 2-UInt(FPMR.LSCALE), before being destructively added without intermediate rounding to the corresponding single-precision elements of the destination vector.
The 8-bit floating-point groups within the second source vector are specified using an immediate index which selects the same group position within each 128-bit vector segment. The 8-bit floating-point encoding format for the elements of the first source vector and the second source vector is selected by FPMR.F8S1 and FPMR.F8S2 respectively.
This instruction is unpredicated.
| 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 | 1 | 1 | i2 | Zm | 0 | 1 | 0 | 0 | 0 | 1 | Zn | Zda | |||||||||||
if !HaveSVE2FP8DOT4() then UNDEFINED; integer n = UInt(Zn); integer m = UInt(Zm); integer da = UInt(Zda); integer index = UInt(i2);
| <Zda> |
Is the name of the third source and destination scalable vector register, encoded in the "Zda" field. |
| <Zn> |
Is the name of the first source scalable vector register, encoded in the "Zn" field. |
| <Zm> |
Is the name of the second source scalable vector register Z0-Z7, encoded in the "Zm" field. |
| <imm> |
Is the immediate index of a 32-bit group of four 8-bit values within each 128-bit vector segment, in the range 0 to 3, encoded in the "i2" field. |
CheckFPMREnabled(); if IsFeatureImplemented(FEAT_FP8DOT4) then CheckSVEEnabled(); else CheckStreamingSVEEnabled(); constant integer VL = CurrentVL; constant integer elements = VL DIV 32; constant integer eltspersegment = 128 DIV 32; bits(VL) operand1 = Z[n, VL]; bits(VL) operand2 = Z[m, VL]; bits(VL) operand3 = Z[da, VL]; bits(VL) result; for e = 0 to elements-1 integer segmentbase = e - (e MOD eltspersegment); integer s = segmentbase + index; bits(32) op1 = Elem[operand1, e, 32]; bits(32) op2 = Elem[operand2, s, 32]; bits(32) sum = Elem[operand3, e, 32]; sum = FP8DotAddFP(sum, op1, op2, FPCR, FPMR); Elem[result, e, 32] = sum; Z[da, VL] = result;
This instruction might be immediately preceded in program order by a MOVPRFX instruction. The MOVPRFX instruction must conform to all of the following requirements, otherwise the behavior of the MOVPRFX and this instruction is unpredictable:
Internal version only: aarchmrs v2023-12_rel, pseudocode v2023-12_rel, sve v2023-12_rel ; Build timestamp: 2023-12-15T16:46
Copyright © 2010-2023 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.