Emulated float128 or double-double arithmetic. A floating point expansion with 31 accurate decimal digits.
