English

Summing $\mu(n)$: a faster elementary algorithm

Number Theory 2022-03-01 v4

Abstract

We present a new elementary algorithm that takes time  Oϵ(x35(logx)35+ϵ)  and  space  O(x310(logx)1310) \mathrm{time} \ \ O_\epsilon\left(x^{\frac{3}{5}} (\log x)^{\frac{3}{5}+\epsilon} \right) \ \ \mathrm{and}\ \ \mathrm{space} \ \ O\left(x^{\frac{3}{10}} (\log x)^{\frac{13}{10}} \right) for computing M(x)=nxμ(n),M(x) = \sum_{n \leq x} \mu(n), where μ(n)\mu(n) is the M\"{o}bius function. This is the first improvement in the exponent of xx for an elementary algorithm since 1985. We also show that it is possible to reduce space consumption to O(x1/5(logx)5/3)O(x^{1/5} (\log x)^{5/3}) by the use of (Helfgott, 2020; arxiv.org:1712.09130), at the cost of letting time rise to the order of x3/5(logx)x^{3/5} (\log x).

Keywords

Cite

@article{arxiv.2101.08773,
  title  = {Summing $\mu(n)$: a faster elementary algorithm},
  author = {Harald A. Helfgott and Lola Thompson},
  journal= {arXiv preprint arXiv:2101.08773},
  year   = {2022}
}

Comments

43 pages, 2 figures. Fourth version: streamlined abstract, rewritten introduction

R2 v1 2026-06-23T22:24:03.333Z