Tooling: 5.52x faster complete PR886 bit-frame replay (no new kappa)

[RossRichardson]

Change and mathematical scope

PR886's historical bit-frame audit rebuilds a modular elimination for each candidate live line. This opt-in adapter builds one echelon basis per closure and keeps the full checker, exact interval comparisons, enumeration and receipts. Verification tooling; no new kappa. Frozen sources, defaults, the selected record and upstream/ remain unchanged.

Five alternating same-input pairs on macOS arm64 / CPython 3.12.12 measured complete three-case suite medians 32.61348 s → 5.90754 s (5.52066x). All samples, including slower pairs on the shared host, are retained. This is a historical verifier improvement, not a selected-replay or multiplication-runtime claim.

Task and overlap

• Base: 2de86c77b4127d2f01f504e0c7c5e4db737caa3c; seed 93; replay-runtime, refined to PR886 closure evaluation. Task and source pins.
• Uses the integrated PR886 refinement, with exact source Git blobs/SHA256 pins for the checker, PR589/555 words and interval engine. Submitted PR886 head: 881f72ec596a5c6d7aba801153792292ecc39737; maintainer refinement: 3b83e4f14825ce1ae038a8c8159f130646991086.
• Tooling: accelerate exact historical frame containment #1459/Tooling: exact annihilator containment for historical frame audits #1461 use related exact containment ideas in different rational-frame checkers. Tooling: 1.45x faster complete R439 omission replay #1433 batches a different Boolean omission audit. Scoped negative result: no relocation or Steinberg rewrite with exact reframing raises PR589's bit c (no new kappa) #933/Scoped negative result: PR589's bit word is exactly transport-locally optimal (no new kappa) #1345 investigate transport search, rather than this closure adapter. No duplicate adapter appeared in the inspected PR/Discussion records; search coverage and limits are recorded in the task card.

Validation

python3.12 scripts/research_package.py --package research/bit-frame-closure-basis

Expected: PASS_BIT_FRAME_CLOSURE_BASIS and PASS_REGISTERED_RESEARCH_PACKAGE.

• Complete reference/candidate comparisons: 6,477 closure entries, 1,156 assignments, all three complete receipts; degraded PR589 still has 37 improving components and recovers the tuned profile.
• Nine control groups, including 1,100 exact Fraction-oracle cases, wrong-membership/lost-pivot mutants, unsafe modular aliasing, source corruption, disabled assertions and stale benchmark evidence.
• Clean read-only minimal export passes the full replay on Python 3.14.6 with an optimized parent and assertion-enabled children.
• make selected-record-check passes. All five consumed blobs and selected-record bytes also match inspected main a18b4b0136d647529971e9fcb99fa83f5b9a0b81.

Proof and scope and validation receipt. No full historical suite, selected cold replay, native/physical supplier admission, Lean build or complete multiplication theorem is claimed. Hosted CI is separate from local evidence.

Credit

Seth Burkart's PR886; Douglas Colkitt's exact integration refinement; DreamingOfClouds' PR589 on eumemic's PR555 and its credited lineage. Related containment work by CubeJerry and omission tooling by Sagar Thakrar are acknowledged. New material by Ross Richardson with substantial OpenAI Codex assistance; Apache-2.0. No subagents were used.

Open next questions

Profile a pinned PR933 or PR1345 checker before attempting reuse. Its mutable-state and rank contracts need separate validation; this PR establishes no result for those workloads.

[Mister-Meeseeks]

Batch review at 0669abfe0d82b7150f3f215dad6b4eaf324e2ce7.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Opt-in one-echelon-basis closure preserves the entire PR886 fixed-word audit and exact intervals, with Hadamard guard, immutable per-word input, unchanged assertions and full differential inventories. Independent Fraction controls; author benchmark is workload-specific.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.