pint.pulsar_mjd.two_sum

pint.pulsar_mjd.two_sum(a, b)

Add a and b exactly, returning the result as two float64s. The first is the approximate sum (with some floating point error) and the second is the error of the float64 sum.

Using the procedure of Shewchuk, 1997, Discrete & Computational Geometry 18(3):305-363 http://www.cs.berkeley.edu/~jrs/papers/robustr.pdf

Returns:

sum, err – Approximate sum of a + b and the exact floating point error

Return type:

float64