BFSCALE (multiple and single vector)

Multi-vector BFloat16 adjust exponent by vector

This instruction multiplies the BFloat16 elements of the two or four first source vectors by 2.0 to the power of the signed integer values in the corresponding elements of the second source vector and destructively places the results in the corresponding elements of the two or four first source vectors.

This instruction follows SME2 non-widening BFloat16 numerical behaviors corresponding to instructions that place their results in two or four SVE Z vectors.

This instruction is unpredicated.

It has encodings from 2 classes: Two registers and Four registers

Two registers
(FEAT_SME2 && FEAT_SVE_BFSCALE)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME2) || !IsFeatureImplemented(FEAT_SVE_BFSCALE) then EndOfDecode(Decode_UNDEF); end; let dn : integer = UInt(Zdn::'0'); let m : integer = UInt('0'::Zm); let nreg : integer{} = 2;

Four registers
(FEAT_SME2 && FEAT_SVE_BFSCALE)

Decode for this encoding

if !IsFeatureImplemented(FEAT_SME2) || !IsFeatureImplemented(FEAT_SVE_BFSCALE) then EndOfDecode(Decode_UNDEF); end; let dn : integer = UInt(Zdn::'00'); let m : integer = UInt('0'::Zm); let nreg : integer{} = 4;