Decoding stage supports FMMACC_B, but it seems like that the mac_float module does not support size_8, according to the information provided.
