Making flexible floating data type

[nmizukami]

Erik suggested that at least output data type need to be either float or double, which can be specified by the user. This is good practice and double precision is needed for testing (e.g., the comparison with test case - bit-by-bit comparison).