There is a newer version of the record available.

Published September 20, 2026 | Version v1

The Atlas of Maximal Gaps: Exact Covering Enumeration for Primorial Sieves

Description

We introduce a framework that converts the search for maximal gaps between integers coprime to a primorial p#—and for maximal deserts between twin-pair centers—into a finite enumeration over the phases of a fixed layer sieve. An atlas is the set of admissible (phase, covering assignment) pairs; we prove that the gap (resp. desert) positions of each length are exactly the CRT lift classes of the atlas entries, with a smallest-prime refinement giving an explicit bijection. Consequences are proved without computation: a characterization of the Jacobsthal function g(p#) and of the maximal twin-desert width W(p) as the largest feasible atlas width; first occurrences as minimal CRT lifts; a mirror involution pairing all positions; a capacity pre-sieve; and a complete defect stratification of phases. A Symmetry Theorem characterizes the palindromic structure of gap kill-patterns in terms of center divisibility. All values g(p#) and W(p) are certified by independent sieve cross-checks for the Jacobsthal values g(p#), 13 ≤ p ≤ 37, and for the twin-desert widths W(p), 13 ≤ p ≤ 31; W(37) ≥ 462 is certified with maximality provisional.

A central distinction is maintained throughout: the Jacobsthal function g(p#) is a property of the primorial period and can exceed 2(p−2); the location-sensitive prime gap bound 2(⌊√x⌋ − 2) is a separate conjecture bounding the prime gap surrounding each position x ≥ 25, verified for every x ≤ 41² by direct computation, and capable of being smaller than the global Jacobsthal value g(p#) at some positions x. The atlas reduces this conjecture to an explicit finite statement at each prime and locates precisely the open difficulty. The certified tables accompany submissions of new integer sequences to the OEIS. The enumeration cost grows with covering-set size; the present implementation reaches its practical limit at p = 41, well inside the range of record computations—the contribution is structural (positions, twin deserts, mirror pairing, the Symmetry Theorem), not record-setting.

Files

atlas_changelist.md

Files (493.3 kB)

Name Size Download all
md5:db8021116e7e36fba7e5cb9256f54445
10.4 kB Preview Download
md5:ba0c12624a81d1fc94a5a0f8389de17d
6.1 kB Download
md5:1abbcd801e3353200f7bcabbc1e8e7ce
224.7 kB Preview Download
md5:f4b3ccd52bdc7b4e2f5efeadeafc3d39
38.9 kB Download
md5:444fb089ec29f4e2435be3ec099adf98
548 Bytes Preview Download
md5:c026a18145ef81415e46209503037b79
690 Bytes Preview Download
md5:ac07674d274f7c6e5c8f3cb66b4dd175
84.9 kB Preview Download
md5:048cb12dd309b3d1dad756e0f0edf592
61.1 kB Preview Download
md5:c7bc6059eb9c9830fa6c98b5d2e53903
63.0 kB Preview Download
md5:3e363c3ed290d16c1eafb51d971ae7fb
2.9 kB Preview Download