The DCOSH function accepts a number and returns the double-precision value of the hyperbolic cosine of that number. Depending on the value of the argument x, DCOSH(x) is computed either directly from the definition by using the DEXP function or by a polynomial approximation.

The argument is reduced to the first quadrant by the method used in calculating the single-precision COS. The calculation then uses the same approximation as is used for DSIN and then uses the following identity:

DCOS(x) = DSIN(pi/2 - x)