結果 : expand single precision and double precision floating point representation