Rigorous high-precision computation of the Hurwitz zeta function and its derivatives (Q2351480): Difference between revisions

The author employs the Euler-Maclaurin summation formula complete with rigorous error bounds to evaluate the Hurwitz zeta function \(\zeta(s,a)\) and its derivatives with respect to \(s\), to extremely high precision (hundreds or thousands of digits). A detailed discussion is given of the various algorithmic techniques employed to reduce the storage space and computing time for such high-precision computations. An application of the above algorithm is made to produce record computations of (i) the first zero of the Riemann zeta function \(\zeta(s)\) to 303000 digits, (ii) the Keiper-Li coefficients \(\lambda_n\) for \(n\leq 10^5\) and (iii) the Stieltjes constants \(\gamma_n(a)\) that appear in the Taylor series expansion of \(\zeta(s,a)\) about the point \(s=1\), also for \(n\leq 10^5\). Both sets of constants in (ii) and (iii) are computed to several thousand digits.