Support–Prime Dual Frontier

Support–Prime Dual Frontier v0.4 — Expands the dual witness method to four support radii, all of which are blocked by safe lower bounds (2.62 → 1.09), while catching a methodological trap: coarse axis grids generate false escapes. The prime enumeration cost at R=16 has reached 79 trillion terms.

RH_CLAIM = False — The files in the package self-report "not a proof of RH, not a disproof of RH, nor an interval certificate for the continuous analytic problem", reproduced here as-is. The safe lower bounds for all four radii are strictly greater than 1, and all 12 witnesses are floating-reconstructed and remain PSD.

Connections

This package generalizes the dual witness method of the previous package to multiple support radii, quoting its decisions as-is.

"Increasing R is no longer the primary direction. The next node co-designs axis notches, dictionary, and cover, and requires any α<1 result to undergo a dense-axis audit." — Excerpted from this package's RESULTS.md § Decision.

Previous Node · v0.3Axis-Target Dual Obstruction

Next Node · v0.5Axis Notch and Cover Co-design

Technical Notes

Main Document — Uniform frontier scanning, joint dual optimization of 288 sub-rectangles, coarse-axis grid false escape diagnostics, actual prime enumeration, and spectral gap steering decisions.

Result Data

Safe lower bound table for four radii, coarse grid false escape comparison, uniform frontier summary, and prime costs, from RESULTS.md within the package.

Trust Boundaries

From TRUST_BOUNDARY.md within the package.

Files

The large-scale spectral scan data (uniform_frontier.json 186 KB) and prime histograms (prime_histograms.npz) are not listed individually; they are all in the full zip file.

Download Complete Package (176.7 KB)

sha256 dc8240726dd87779e367c40b13d12060ba7cfeda6741dfafdedc790859366363