A fast algorithm to compute the Ramanujan-Deninger gamma-function and some number-theoretic applications

TL;DR

This paper presents a fast algorithm for computing the Ramanujan-Deninger gamma function and its derivatives, enabling extensive numerical analysis of Euler-Kronecker constants and related number-theoretic quantities for large primes.

Contribution

The authors develop a new efficient algorithm for the Ramanujan-Deninger gamma function, allowing large-scale numerical investigations of key number-theoretic constants.

Findings

Computed Euler-Kronecker constants for large primes, including a new negative value.

Extended numerical results for primes up to 10^7, confirming positivity of constants.

Established bounds for the maximum of derivatives of Dirichlet L-functions.

Abstract

We introduce a fast algorithm to compute the Ramanujan-Deninger gamma function and its logarithmic derivative at positive values. Such an algorithm allows us to greatly extend the numerical investigations about the Euler-Kronecker constants $\mathfrak{G}_q$, $\mathfrak{G}_q^+$ and $M_q=\max_{\chi\ne \chi_0} \vert L^\prime/L(1,\chi)\vert$, where $q$ is an odd prime, $\chi$ runs over the primitive Dirichlet characters $\bmod\ q$, $\chi_0$ is the trivial Dirichlet character $\bmod\ q$ and $L(s,\chi)$ is the Dirichlet $L$-function associated to $\chi$. Using such algorithms we obtained that $\mathfrak{G}_{50 040 955 631} =-0.16595399\dotsc$ and $\mathfrak{G}_{50 040 955 631}^+ =13.89764738\dotsc$ thus getting a new negative value for $\mathfrak{G}_q$. Moreover we also computed $\mathfrak{G}_q$, $\mathfrak{G}_q^+$ and $M_q$ for every odd prime $q$, $10^6< q\le 10^7$, thus extending previous results. As a consequence we obtain that both $\mathfrak{G}_q$ and $\mathfrak{G}_q^+$ are positive for every odd prime $q$ up to $10^7$ and that $\frac{17}{20} \log \log q< M_q < \frac{5}{4} \log \log q $ for every odd prime $1531 < q\le 10^7$. In fact the lower bound holds true for $q>13$. The programs used and the results here described are collected at the following address \url{http://www.math.unipd.it/~languasc/Scomp-appl.html}.