Consume retimed staged releases with smaller T360 stock

research/staged-donor-release/CONCURRENT-WORK.md

@@ -0,0 +1,34 @@
+# Concurrent work found at the final checkpoint
+
+The final complete overlap refresh (374 open PR bodies and 2464 Discussion
+records, zero pagination warnings, 2026-10-18 13:33 UTC) found concurrent work
+not returned by the earlier narrower literal-term queries. No priority claim
+is made. Canonical main remains 20ecf2891d80b5acf18d980b2164fb44ea7faa90.
+
+- [PR3642](https://github.com/CrocSwap/integer-mult-bounds/pull/3642), current
+  head 044d24f12e74f228dd462c07ae2d7f3ea35e2be9, created 12:30 UTC,
+  reports the same two commuting release retimings with a T530 allocation.
+  Its current conditional endpoint is .001730394391813963415120 minus the
+  same backoffs. Its retiming contribution deserves explicit credit.
+- [PR3653](https://github.com/CrocSwap/integer-mult-bounds/pull/3653), current
+  head 607a09ff441c79b215a8127a81f4cc70021d0d9b, created 13:08:59 UTC,
+  composes those suppliers with T488 and reports exactly this package's
+  conditional endpoint .001730418899766071163804−10^-46−10^-70.
+  Its declared counts satisfy h488/488=h360/360 for every rank; cross-products
+  were checked exactly. We have not independently replayed its release.
+
+This T360 package thus **matches the concurrently reported retimed endpoint
+with smaller stock**: 323069 instead of 437757 physical families, 3240 instead
+of 4392 calls, and 1321920 instead of 1791936 literal helper roles. That stock
+reduction changes finite constants, not the normalized exponent. Its exact
+local construction and conditional finite release remain a strict improvement
+over the explicitly pinned PR3638/PR3633 predecessor. It is not a further
+frontier increase over PR3653. No new selected-frontier value is asserted.
+
+The local result was independently investigated during this campaign, but the
+final draft is framed as T360 composition plus independent source/evidence
+checks, with the concurrent mechanism and endpoint credited. Frozen scientific
+receipts, comparator pins and the original review at a3bdc591 are preserved.
+The public PR bodies/head/timestamps read at this checkpoint are retained under
+evidence/concurrent. They are provisional external evidence, not scientific
+admission. None of these PRs was altered, reviewed wholesale or closed here.

research/staged-donor-release/LEAD1-CONTROLS.json

@@ -0,0 +1,76 @@
+{
+  "status": "PASS_SOURCE_BOUND_ENDPOINT_CONTROLS",
+  "transport_sha256": "9818e08b819310b40f1441b89be8c4369584a5b3fa5dfe9b86b1df9a7f5d6a76",
+  "checks": [
+    {
+      "provider": "base",
+      "control": "valid",
+      "exit_code": 0,
+      "endpoint_rejected": false,
+      "log_sha256": "912da5c69252a47dd82b3dd02a75f69519de43d05ce37fd63dc455315e295b7f"
+    },
+    {
+      "provider": "base",
+      "control": "omit-release-644",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "fe440dff56d88ffb1254befeb9b4d0034da4e471e63f93c0a0eaf1b472ae7b70"
+    },
+    {
+      "provider": "base",
+      "control": "wrong-sign-644",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "fe440dff56d88ffb1254befeb9b4d0034da4e471e63f93c0a0eaf1b472ae7b70"
+    },
+    {
+      "provider": "base",
+      "control": "omit-release-660",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "7415a2086c69858488094fc3102766a0c3f53ecd637852ac5f0612ff3a6960aa"
+    },
+    {
+      "provider": "base",
+      "control": "wrong-sign-660",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "7415a2086c69858488094fc3102766a0c3f53ecd637852ac5f0612ff3a6960aa"
+    },
+    {
+      "provider": "extra",
+      "control": "valid",
+      "exit_code": 0,
+      "endpoint_rejected": false,
+      "log_sha256": "08199ef6bbb9959ae9de501a2bf8c4c0c8c3291334f8ec2e05e6e688ed04ef7a"
+    },
+    {
+      "provider": "extra",
+      "control": "omit-release-644",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "fe440dff56d88ffb1254befeb9b4d0034da4e471e63f93c0a0eaf1b472ae7b70"
+    },
+    {
+      "provider": "extra",
+      "control": "wrong-sign-644",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "fe440dff56d88ffb1254befeb9b4d0034da4e471e63f93c0a0eaf1b472ae7b70"
+    },
+    {
+      "provider": "extra",
+      "control": "omit-release-660",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "7415a2086c69858488094fc3102766a0c3f53ecd637852ac5f0612ff3a6960aa"
+    },
+    {
+      "provider": "extra",
+      "control": "wrong-sign-660",
+      "exit_code": 1,
+      "endpoint_rejected": true,
+      "log_sha256": "7415a2086c69858488094fc3102766a0c3f53ecd637852ac5f0612ff3a6960aa"
+    }
+  ]
+}

research/staged-donor-release/LEAD1-PRICE.json

@@ -0,0 +1,36 @@
+{
+  "dual": {
+    "patterns": 2551,
+    "symbolic_tight": 12,
+    "minimum_strict_lower": "4053468025645212954130191300621690725601209608148213926713/1000000000000000000000000000000000000000000000000000000000000",
+    "scope": "Old capacity and pattern alphabet, evaluated at stated K; not a dual for arbitrary changed widths."
+  },
+  "status": "PASS_EXACT_LOCAL_AND_FIXED_ALLOCATION_PRICE",
+  "k": "173009636352862353423/100000000000000000000000",
+  "local_delta": {
+    "lower": "-3299714488759840734997518419826012212910695186919289400037/500000000000000000000000000000000000000000000000000000000000",
+    "upper": "-6599428977519681469995036839652024425821390373838578800073/1000000000000000000000000000000000000000000000000000000000000"
+  },
+  "width8_to7_shadow_times20": {
+    "lower": "246935045936431659522948237548685897943499745459984726817273/12500000000000000000000000000000000000000000000000000000000",
+    "upper": "19754803674914532761835859003894871835479979636798778145381841/1000000000000000000000000000000000000000000000000000000000000"
+  },
+  "prior_staged_net_times10": {
+    "lower": "-65268039542568976533836655295640969913942214558880321551831/1000000000000000000000000000000000000000000000000000000000000",
+    "upper": "-6526803954256897653383665529564096991394221455888032155183/100000000000000000000000000000000000000000000000000000000000"
+  },
+  "profile": {
+    "1": 7980120,
+    "2": 2345760,
+    "3": 755730,
+    "4": 330480,
+    "5": 69930,
+    "8": 25920,
+    "48": 137790
+  },
+  "root_interval": [
+    "0.001730418899766071163804",
+    "0.001730418899766071163805"
+  ],
+  "scope": "Exact finite paid profile using unchanged T360 idle and private stock. No native, physical ordinary/outer admission follows from these inequalities."
+}

research/staged-donor-release/PR-DRAFT.md

@@ -0,0 +1,51 @@
+# Consume retimed staged releases with smaller T360 stock
+
+## Change and mathematical scope
+
+Compose the two commuting staged-release retimings with the literal T360 allocation. The complete conditional finite release supports **κ = 0.001730418899766071163804 − 10^-46 − 10^-70**, matching the endpoint reported by [PR3653](https://github.com/CrocSwap/integer-mult-bounds/pull/3653) with **323,069 instead of 437,757 physical families**, 3,240 instead of 4,392 calls, and 1,321,920 instead of 1,791,936 helper roles. The normalized profiles are identical; smaller stock changes finite constants. No priority or further frontier gain over PR3653 is claimed.
+
+In PR3605's words, move `640 += 2*644` and `656 -= 660` to immediately before the pivots' first producer updates. Exact commutation preserves the full arbitrary-dirty operator, while each donor pays H3 instead of H1+H2. Both providers have local profile `{1:2437,2:724,3:226,4:102,5:20,8:8}`. The complete paid profile is `{1:7980120,2:2345760,3:755730,4:330480,5:69930,8:25920,48:137790}` across 322,560 live families, with 509 additional private families. The invoice includes both providers, all literal roles, 91 zero buffers, and the charged 2,048-bit prefix reserve.
+
+All-size/compiler, uniform source/hosting, recursive precision/tape/recovery, bootstrap, analytic and BIT hypotheses remain inherited. This is a conditional finite construction, not an unconditional multiplication theorem or a selected-frontier update.
+
+## Task and overlap
+
+Explicitly assigned staged-release consumer work; task `independent-moment`, **changed scope**: change and fully price an actual word, rather than only audit an unchanged moment. See `research/staged-donor-release/TASK.json` for the original card and scope rationale.
+
+Base: `20ecf2891d80b5acf18d980b2164fb44ea7faa90`. Exact predecessor inputs are retained with source hashes and notices:
+
+- PR3605 staged words: `3efbb5de06a49ffeea6684482f2d3fe48cdca3fb`.
+- PR3638 T360 allocation: `d62e1acd7518a231440383f0d0dddf54d6edf8c1`.
+- PR3633 packing comparison: `72c268f113368ea0387d188225e3f81ba6d6ef73`.
+
+Concurrent [PR3642](https://github.com/CrocSwap/integer-mult-bounds/pull/3642) reports the same retiming mechanism; PR3653 composes it with T488. This independently constructed package contributes the smaller T360 consumer and its source-bound validation. The strict gain of `0.000000322536237447629574` is relative only to the pinned PR3638/PR3633 predecessor on the common release grid. The final full overlap search covered 374 open PR bodies and 2,464 Discussion records; the related PR heads were rechecked during preparation. Coverage is not a novelty proof. See `CONCURRENT-WORK.md` for exact heads and comparisons.
+
+## Validation
+
+- Both changed providers passed fresh five-stage native replay: 632 arbitrary columns in both orientations, phase/warp recipes, and actual omission/sign controls.
+- The complete original release passed all strict finite margins, with the predecessor edit chain retained and a separately checked retiming stage. The original run explicitly reused authenticated unchanged native/parent evidence and preparatory discovery after a declared-stage binding repair.
+- Exact dual calibration, integral T360 fill, independent frame/invoice checks, and owner-transfer/source-closure checks passed. Fresh Astra medium review approved the scoped scientific result at `a3bdc591920a4b9003d8442d1fcca24218153661`; a follow-up review approved the concurrent-work framing. These are independent AI reviews, not human or formal-kernel approval.
+- Bounded registered replay, committed-byte comparison and Git-free committed export replay pass for the prepared package. The new full-wrapper evidence is separate from the frozen original scientific evidence.
+
+Reproduce the bounded package with Python 3.12 (standard library only):
+
+```sh
+python3.12 -B scripts/research_package.py --package research/staged-donor-release --check-only
+python3.12 -B scripts/research_package.py --package research/staged-donor-release
+```
+
+Full replay: Python 3.12.8, NumPy 1.26.4, SciPy 1.15.3, SymPy 1.14.0, mpmath 1.3.0; assertions enabled, one heavy replay at a time:
+
+```sh
+python3.12 -B research/staged-donor-release/full_replay.py --work /tmp/fresh-staged-release
+```
+
+The default full wrapper **passed end to end**, including both fresh changed native providers, fresh preparation and all seven complete-admission stages. The wrapper and scientific inputs match commit `74afc9ff5cb437d3112c6279216d10250775cd98` and are unchanged in this preparation update. Raw receipts, logs, environment versions and hashes are retained in `research/staged-donor-release/evidence/full-wrapper/`; the bounded verifier authenticates them and rejects four targeted receipt inconsistencies. This later run retains explicitly authenticated unchanged parent/native ancestry; it is execution evidence subsequent to the cited independent AI review. A no-reuse retained-parent cold run and hosted CI have not been run. Source-history preflight refused the shallow checkout; the self-contained archive instead reconstructs all required source bytes without foreign Git objects. No Lean or Docker was run.
+
+## Credit
+
+Credit Ross Richardson/PR3605, roger-sato's concurrent PR3642, Rohan Arun's PR3653, Pierre-Loic Doulcet's PR3633, PR3638's T360 allocation, dcharb78/PR3133, Marius-Paul Dumitrean/PR2979, Jeremy Kothe, and the retained native, owner, block-norm and outer authors. Exact source notices remain in the package. Apache-2.0; substantial OpenAI Codex assistance and independent AI review.
+
+## Open next questions
+
+See the package README and structured `TASK.json` questions. A useful next construction would identify one compensated later release or changed joint consumer, prove its arbitrary-dirty identity, and price its complete invoice before synthesis. The inherited uniform/all-size premises remain separate open obligations.

research/staged-donor-release/PROOF.md

@@ -0,0 +1,104 @@
+# Lead 1: delayed donor release on the staged words
+
+LOCAL-ONLY complete conditional finite release; independent terminal approval recorded at a3bdc591920a4b9003d8442d1fcca24218153661. No selected-frontier change.
+
+The changed producer neighborhood is helper644/donor640 and its symmetric
+helper660/donor656. Unlike another endpoint extraction on PR3133, the input
+is PR3605's actual staged word, authenticated by the compressed pins in
+SOURCE-PINS.json. Move B591 `640 += 2*644` to just before old B1045;
+move B641 `656 -= 660` to just before old B1085. The other delayed releases
+(into496/520) stay at their staged positions. Every warp table entry retains
+its original identity; only its B cut is reindexed. Frames and source symbols
+of every other instruction are retained; full inverse cleanup is regenerated.
+
+Between each old/new cut the pivot is never written, the donor is never read,
+and neither crosses a native warp. Thus every intervening elementary matrix
+commutes with the release shear. The complete operator is exactly unchanged
+on arbitrary independent dirty states, including both inverse-transpose words.
+The same prefix/suffix product equality preserves all632 inputs and91 private
+zero buffers. Native replay independently checks every column in both
+orientations and both providers. Omit/sign mutants regenerate cleanup and
+must fail the actual endpoint, not merely a declared record comparison.
+
+The late release uses the pivot's first producer frame (rank4). Its donor now
+jumps directly from its rank1 entrance to rank4 rather than through rank2.
+Per donor delta is H3-H2-H1. For0<k<1, strict subadditivity of x^(1-k) gives H3<H1+H2, independently of numerical root extraction. Exact local profile is
+{1:2437,2:724,3:226,4:102,5:20,8:8}; inventory stays
+{4:26,5:2,6:18,7:52,8:97,9:213}. Widths,408 ordinary roles,8 centres,
+91 buffers,2 work families and the T360 chronology/owners stay fixed.
+Exact all-pattern dual check at the comparator K has2551 patterns,12 symbolic
+equalities; the unchanged RHS has zero dual delta. An actual existing primal
+is retained, so this negative local delta is an upper bound using that fill,
+not just a favorable dual screen. The resulting finite full-profile root is
+in [0.001730418899766071163804,0.001730418899766071163805]. This is a
+profile endpoint, now bound to the completed original finite outer release; it is not a complete multiplication theorem.
+
+## Obligation ledger
+
+Proved/finite checked: literal commuting identity, full local frame chains,
+actual entrances and residual widths, complete local bill, both dirty endpoint
+orientations, native five-stage phases and recipes, old exact dual calibration.
+Constructive inherited upper bound: same integral T360 fill/owner inventory and
+unchanged private stock. Its source/native rebinding and all original finite admission gates passed. Terminal receipts and the explicit preparatory reuse coverage are in evidence/complete.
+Unknown/inherited: all-size compiler, uniform source/hosting and refined primes,
+rounded recursive precision/tape/recovery, bootstrap and analytic/bit H1-H5.
+The complete finite release remains conditional on these predecessor premises.
+PR3562's ordinary-backend route is not a universal cap; its distinct physical
+root/movement hypotheses are lead4 and cannot be silently inserted here.
+
+## Numeric fields and historical template prose
+
+The strict supplier/VERIFY/ASSEMBLY records contain the separate coarse online
+saving `0.001727429722725848`. The claimed conditional endpoint comes from the
+certified `complex_lower=0.001730418899766071163804` and the modern joint outer
+ledger, subtracting both exact release backoffs. These are different layers,
+not interchangeable κ fields. The stored OUTER-CLAIM `branch` string retains
+historical 3000-digit/e-3080 prose. The actual `outer_refinement`, `retuned_slacks`,
+numeric ledger and release use endpoint precision24 and the e-104 free scale.
+The inherited template prose is preserved as source history; it is not the
+numeric parameter record for this release. All applicable numeric binding and
+strict-margin checks pass.
+
+## Checks and limits
+
+`python3 -B research/staged-donor-release/delay_release.py --out /tmp/fresh-delayed-words`
+`python3 -B research/staged-donor-release/price.py`
+`/tmp/three-kappa-env/bin/python -B research/staged-donor-release/controls.py --transport TRANSPORT`
+Two new five-stage native runs: Python3.12.8,numpy1.26.4,scipy1.15.3,
+sympy1.14.0,mpmath1.3.0; assertions enabled, one job.
+Original stages and exact source inventories are in evidence/native-*.
+No Lean, Docker or complete no-reuse cold CLI was run. The first full attempt correctly rejected a missing declared retiming stage in the source reconstruction. V2 preserves the PR3605 intermediate and all original reconstruction controls, appends the new stage plus an independent reconstruction, and reruns the complete release. Preparatory discovery/outer arithmetic is explicitly reused under exact changed-source coverage; both actual native words and every scientific native/global engine are unchanged. The portable archive retains the exact V1 differing source bytes and manifest as well as V2, so this reuse has a self-contained source closure. The APFS copy adapter
+only clones regular file bytes; all scientific validators remain unchanged.
+
+The later default `full_replay.py` invocation also passed end to end, with both
+changed native providers fresh and all seven admission stages regenerated. Its
+receipts and precise reuse scope are retained separately under
+[evidence/full-wrapper](evidence/full-wrapper/README.md). This later execution
+does not alter the original scientific source, receipts or independent review.
+
+## Comparator, scope and credit
+
+Canonical main20ecf2891d80b5acf18d980b2164fb44ea7faa90, priority revision6;
+selected conditional kappa0.001672660197076918. Live PR3638 atd62e1acd7 and
+PR3633 at72c268f11 tie at0.001730096363528623534230 minus release backoffs.
+These are pinned historical comparisons. Concurrent PR3642 reports the same
+retimings and PR3653 reports the same normalized endpoint at T488; this T360
+consumer has smaller finite stock. See [CONCURRENT-WORK.md](CONCURRENT-WORK.md).
+The initial overlap scan fully paginated348 open PRs and2454 discussions;
+that is coverage of bodies/comments, not all source diffs or a novelty proof.
+Source archive ledger intake remains partially reviewed, not scientific approval.
+
+Credit: Ross Richardson/PR3605 staged entrance; dcharb78/PR3133 coupled fold;
+Marius-Paul Dumitrean/PR2979; Jeremy Kothe's word/restoration lineage;
+PR3633/3630 exact fractional packing; PR3638 T360 allocation and source-bound
+reuse; Tekalign/Yiyang Lu and retained native, owner, block-norm and outer
+authors. SOURCE-PINS preserves original bytes and source-specific notices.
+New research and code prepared with substantial OpenAI Codex assistance.
+Independent AI review is retained under reviews/lead1; no human or kernel review.
+
+## Open next questions
+
+The complete original admission retains its strict finite margins after the literal source/profile rebinding. Fresh Astra medium terminal review explicitly approves this scope at a3bdc591920a4b9003d8442d1fcca24218153661.
+Later compensated releases after pivot updates and jointly changed interior
+consumers remain outside the two commuting shifts proved here. No general
+mechanism exhaustion or claim of optimality is intended.