Repository navigation
Conversation
|
Thank you @eumemic. This contribution is now recorded on main in the scoped reconciliation review and contributor credits. Contribution: Restrict to 22 and replay new supports and coefficients. Review and limits: Exact-head CI evidence; old geometry is not assumed unchanged. The complete original source, proof/research notes, notices and relevant replay/CI evidence are retained through the pinned source archive. The review distinguishes fresh runs, exact-head historical CI and prior integration evidence. Closing this historical PR as reconciled; its obsolete headline does not replace the current selected bound, and draft/retracted claims retain their explicit qualifications. Reviewed head: |
3 participants
κ ≈ 4.20151003347 × 10⁻⁴.
Restrict the immutable PR117 scalar DAG to its first 22 coordinates, regenerate its physical word, and put it in the paid sequential triple cover from #137. This improves #137 and its slack-tightened successor #138 by about 1.58% under the same inherited multiplication interfaces. All three stages retain completed sequential dirty-transparent cores.
The restriction zeros omitted inputs, contracts unary additions and merges equal supports. The original witness remains pinned; the verifier reconstructs the derived DAG and checks every retained root support. The resulting word has 64,140 live additions, 22,977 virtual scratch roles and 21,274 physical slots after 1,703 compensated reuses. Readout coefficients use exact numerator/38 chunks. Arbitrary operation frames and source gauges are regenerated with the padded-cover objective.
All nine local words, 9,240 width-two complex data finishes and every physical auxiliary tail remain paid. Zero-rank tails retain their scalar adapter charge. The finite router and semantic budget are recomputed using the complete group order. The local scalar guard uses 22,977 virtual roles and 85,577 reserved workspace operations, above the actual 76,778. The independent literal audit counts all 7,075,117 expanded scalar operations per stage, checks exact frame incidences, arbitrary dirty cancellation and the actual reversed chronology.
The complex saving is 210252177/500000000000 = 0.000420504354. The unchanged weighted bit supplier supports 11265104579/25000000000000 = 0.00045060418316, including the full fallback, atom conversion and restored internal rows. Exact assembly uses β = 10⁻¹⁸, η = 10⁻¹⁵ and backoff 10⁻²⁶; the complex side binds.
Validation at current head
2fc6cf5: deterministic fresh restriction/compiler/gauge/reuse regeneration and the full literal reflection audit pass. The finite geometry checks all subspaces and ordered pairs through dimension four and all 1,540 actual ports; padded geometry checks all 764 physical source-frame types. Exact assembly passes 47 strict constraints, seven positive margins and rejection of the next grid point. All 37 corruption controls pass, including in the publication checkout. The complete frozen aggregate passes: original-witness restriction, fresh compiler and independent literal reflection, both padded ledgers, exact assembly, all 37 controls, and an unchanged source snapshot. Generated profiles and receipts reproduce byte for byte. All 45 current-head CI checks pass, including the complete focused verifier on Python 3.11, 3.13 and 3.14.The aggregate reconstructs everything in a temporary copy, requires byte-identical profiles and receipts, and checks that the frozen source closure stays unchanged. The focused workflow covers Python 3.11, 3.13 and 3.14. Concise arguments are in
research/cover-local-reuse/PADDED-GEOMETRY.mdandBIT-PADDED.md.This remains a conditional finite multiplication witness for
T(n) = O(n(log n)^(1−κ)). General Clifford normal forms, common generic bases, uniform weighted local-ring compilation, restored row borrowing, analytic recovery and fixed-tape transfer remain explicit inherited contracts. The simultaneous-core execution proposed in #139 is a separate, explicitly unproved transfer hypothesis and is not used here.Credit: icekylinx with OpenAI GPT-6 Astra/Codex assistance for #130's cover, generalized Clifford theorem and weighted bit interfaces; eumemic with Anthropic Claude assistance for #117's scalar DAG; James Chang/jamesyc for #124's compensated reuse; an664 for completed-core sharing; Avi Eisenberg/ikeboy, Zhihao Chen/jacklightChen, Swapnil Jain, Rohan Arun, Aurel Prosz, RaD and all inherited contributors. Restriction, physical frame/gauge integration and verification were prepared for eumemic with OpenAI Codex assistance. Original licenses, source-specific notices and disclosures are retained and pinned by the manifest.