The DTANH function accepts a number and returns the double-precision hyperbolic tangent of that number. The double-precision hyperbolic tangent is computed either directly from the definition by means of the intrinsic DEXP or by approximation, depending on the value of the argument (in the same manner as TANH).