Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
240 changes: 240 additions & 0 deletions research/z4-paired-frame-restoration-seed56/CERTIFICATE.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,240 @@
{
"assignment": {
"cycle_order": 4,
"os_seed": 56,
"track": "recursive interfaces"
},
"base_commit": "bce54fb3f895cbe5cd4c0eb477c3afeed100190f",
"boundary": {
"children": 5,
"dynamic_program_total": 140,
"eager_ADD": 270,
"eager_NEG": 30,
"eager_total": 300,
"eager_word_sha256": "90bf97554a5a1529e96f128ad93e54aeadbfeffbdd73f3c546ceb4f1e0d91273",
"exhaustive_path_count": 1024,
"exhaustive_path_minimum": 140,
"macro_alphabet": "canonical powers P^u and Q^v using the listed explicit words",
"optimized_ADD": 126,
"optimized_NEG": 14,
"optimized_word_sha256": "3459ba0aca8bd5e598696333a1a5b520d4a0e5353553d685732cf7d49acbd097",
"saved_per_batch": 160,
"states": [
0,
1,
2,
3,
0,
0
],
"transitions": [
{
"P_power": 0,
"Q_power": 1,
"after": 1,
"before": 0
},
{
"P_power": 0,
"Q_power": 1,
"after": 2,
"before": 1
},
{
"P_power": 0,
"Q_power": 1,
"after": 3,
"before": 2
},
{
"P_power": 0,
"Q_power": 1,
"after": 0,
"before": 3
},
{
"P_power": 3,
"Q_power": 0,
"after": 0,
"before": 0
}
]
},
"child": {
"checked_basis_columns": 96,
"distinct_block_checks": [
{
"ADD": 62,
"NEG": 6,
"basis_columns": 32,
"block": 0,
"inverse_word_sha256": "12a83c7b67cc26872a45569df66568dce8d6f47697ae23d08a7731d1c3e957d4",
"word_sha256": "24a9f259f8d8bcd826a8feec7f0dbbf22f10c5fa10e2ed0b5e101ea22bebfe95"
},
{
"ADD": 62,
"NEG": 6,
"basis_columns": 32,
"block": 1,
"inverse_word_sha256": "ed74b775d98f56e0b6a9b6153f5dce0e352d7597dd275e84bdfe7ef91db8282c",
"word_sha256": "33c9708615463d38ac54c283ff87d3755c2a02edfa6535a3fdce7555ed8091ea"
},
{
"ADD": 62,
"NEG": 6,
"basis_columns": 32,
"block": 2,
"inverse_word_sha256": "8f1ba7d29abb2b862f689a3ca7aec86cc9c7c8387aa5d66c905472dbcd6f46b5",
"word_sha256": "2d9232ba4a81a06041bf4057a798aa538985524cbf97b40caca085fcefd309e4"
}
],
"per_call": {
"ADD": 62,
"NEG": 6,
"total": 68
},
"selection_blocks": [
0,
1,
2,
0,
1
]
},
"complete_echo": {
"all_arbitrary_dirty_basis_columns": 56,
"dispatches": {
"eager": 1328,
"optimized": 1008
},
"eager": {
"ADD": 1208,
"NEG": 120,
"total": 1328
},
"eager_word_sha256": "ee0e4afaf3cf2b1dbfc65e2eff076c92dba34fe8aa768fc8b506f9d22382c883",
"internal_C_v_example": {
"all_arbitrary_input_and_scratch_basis_columns": 56,
"eager": {
"ADD": 1262,
"NEG": 126,
"total": 1388
},
"eager_word_sha256": "b63466d5724bf7e60b6fcdb8bca73feaec48788326b43f70804dff1764cc07dc",
"optimized": {
"ADD": 974,
"NEG": 94,
"total": 1068
},
"optimized_word_sha256": "436ab81a02fc6013b6b167937c2294d4547b6f8ffd3888db5569f5c2cedc3131"
},
"layout": {
"output_D": 8,
"source_X": 24,
"total_data_coordinates": 56,
"working_W": 24
},
"map": "(D,X,W)->(D+sum_i R_i X,X,W)",
"optimized": {
"ADD": 920,
"NEG": 88,
"total": 1008
},
"optimized_word_sha256": "ecaf6008f7c81a6c495ce146b971b53d4c3a9b1fea6289645285045796235451",
"prefix_row_l1_peak": {
"eager": 11,
"optimized": 20
},
"reads": {
"eager": 2536,
"optimized": 1928
},
"saved_total_primitives": 320,
"writes": {
"eager": 1328,
"optimized": 1008
}
},
"controls": {
"count": 4,
"names": [
"rotation sign mutation",
"wrong output phase",
"missing terminal restoration",
"forward word substituted for inverse"
]
},
"frame": {
"determinant": 1,
"intertwining_verified": true,
"order": 4,
"output_octets": 1,
"per_register": "two disjoint 4-cycles on one octet",
"phase_costs_per_octet": [
{
"ADD": 0,
"NEG": 0,
"phase": 0,
"total": 0,
"word_sha256": "4f53cda18c2baa0c0354bb5f9a3ecbe5ed12ab4d8e11ba873c2f11161202b945"
},
{
"ADD": 18,
"NEG": 2,
"phase": 1,
"total": 20,
"word_sha256": "652fe5774a3dc187f9df6fd2e578f0270a4a898e76edd1c5e57446ad5281d6fa"
},
{
"ADD": 12,
"NEG": 4,
"phase": 2,
"total": 16,
"word_sha256": "3db669a9548039bbfc034e92b0c3edf8a540d8b147cb95618eddd3e225e13bb4"
},
{
"ADD": 18,
"NEG": 2,
"phase": 3,
"total": 20,
"word_sha256": "278ea7a94adedef7155f32189da6002570d43d68701817674a62494c22d0ea22"
}
],
"readers": [
0,
1,
2,
0,
1
],
"workspace_octets": 3
},
"recursive_interface_cost": {
"complete_root_echo": "2*sum(child vectors) + (48,0) + 2*g(k_v)",
"depth_storage": "56 data coordinates for the root echo plus 24 per active nested internal C_v level",
"eager_comparator": "replace g(k_v) by k_v*(54,6)",
"g(k)": "minimum 4-state recurrence below; vector ADD/NEG invoice retained along chosen path",
"internal_C_v": "2*sum(child vectors) + (48,0) + 2*g(k_v) + (54,6)",
"leaf_forward": {
"ADD": 62,
"NEG": 6
}
},
"schema_version": 1,
"scope": "Exact finite dirty-state interface construction and macro-alphabet optimum for five children; no actual supplier admission, selected profile change, new kappa, or multiplication theorem.",
"selected_context": {
"decimal": "0.001672660197076918",
"kappa": "836330098538459/500000000000000000",
"review": "docs/research/community-batch1099-review.md",
"selected_source_pr": 1094
},
"source_pins": {
"certificates/selected-result.json": "10b9e340fd8e354a37a0c4813c90e5de1dd42044b5f6aca4a3fc5a42b685fa32",
"docs/research/agent-task-catalog.json": "89b5928bc92c81e2647b6d702a7ad85a6795b8f44e2eb267b2ff3145d8f2b2ce",
"docs/research/community-batch1099-review.md": "7e76d2c38e5fce30587835b7251df1430488ae4dd372611ef69a3f03f476b6cb",
"docs/research/current-status.md": "1f59d648763c77d574d18dc59ea8bf62d2ee74aaed3c202c9c8cfa4f901badbd",
"research/community-frontier1094-audit/ASSEMBLY.json": "1122cb3a575f8c788345fc1bf4c108ac7c3250fb81a5988671f5eb04b3ab3045",
"research/community-frontier1094-audit/SELECTED-SOURCE.json": "79b3de5db3dbcb9829789ba58e9e45758926031e2a7d80667a08429d369be53b"
},
"task": "z4-paired-frame-restoration-seed56"
}
35 changes: 35 additions & 0 deletions research/z4-paired-frame-restoration-seed56/CLAIM.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,35 @@
# Seed 56 task selection and overlap record

**Track:** recursive interfaces. **OS seed:** 56. **Base:** `bce54fb3f895cbe5cd4c0eb477c3afeed100190f`. The seed maps to cycle order `2 + (56 mod 3) = 4`.

## Three narrow candidates considered

1. **Selected #1094 T16 same-owner call boundary.** Test whether one concrete repeated owner restore is immediately rebuilt by its next chronological call. The selected review and [heterogeneous owner assembly](../heterogeneous-owner-assembly/README.md) bound the frozen owner catalogue, while recent #1240/#1348 reports quantify selected-word packing headroom and changed coverage-law thresholds. PR #1252's `Z4` is a rank-pair phase polynomial in idle owner repartitioning, not a fourth-order output frame. The task still needed an exact identified two-call transition and a new legal return contract; no such concrete transition was established in this orientation pass.
2. **Signed-reader restoration coalescing.** A specific checkpoint-correct cut problem for fixed dirty readers could reduce restoration. This overlaps current #1199 (diagonal signed readers), #1203 (noncommuting reader restoration) and #1246/#1227 (paid reader-defect and output-sign repair), so it was not selected.
3. **Order-four paired-cycle output-frame boundary.** Extend the already explicit cyclic output-frame mechanism to a genuinely different phase group and gate-cost profile. Current #1287 is order three; its recurrence does not cover the order-four paired cycles, whose phase powers require NEG accounting. Exact-term search found no matching order-four frame task.

## Chosen scope and claim

I chose candidate 3. The task is the five-child dirty echo with three workspace octets, one output octet, and selectors `(0,1,2,0,1)`. The checked macro alphabet is the explicit set of whole-register powers of the paired order-four frame. It compares a complete optimized echo with the eager per-child inverse frame.

Claim posted on [coordination discussion #1163](https://github.com/CrocSwap/integer-mult-bounds/discussions/1163#discussioncomment-18911404) at 2026-10-13 23:35 UTC; expected finish 2026-10-14 00:08 UTC, expiry 01:35 UTC.

## Finding

The certified path is `[0,1,2,3,0,0]`: the five-child boundary costs 140 primitives (`126 ADD, 14 NEG`) versus 300 eager (`270 ADD, 30 NEG`). The complete arbitrary-dirty echo costs 1,008 (`920 ADD, 88 NEG`) versus 1,328 (`1,208 ADD, 120 NEG`), saving 320. The corresponding internal `C_v` map and its literal inverse are also checked over all 56 input/scratch basis columns. The exact 9-ADD/1-NEG quarter-turn word was found during the final cost audit; an earlier three-pure-swap word was replaced before freezing the result. The final prefix row-L1 peak rises 11 to 20.

## Overlap evidence

At 2026-10-13 23:33 UTC, `python3 scripts/find_overlap.py 'order-four cycle output frame Z4 paired cycles' --refresh --limit 12` searched complete snapshots of **839 open PRs** and **89 discussions** and returned zero exact multiword matches. A later offline cached query at 23:44 returned PR #1287 for `cyclic output frames`, the posted seed-56 claim for `paired 4-cycles`, and PR #1252 for `Z4`. Direct inspection confirmed that #1252 uses `Z4` for a rank-pair phase polynomial in idle owner repartitioning, not an order-four output frame. The helper does not search arbitrary PR code diffs or nested package files; this is not a novelty certificate. Direct inspection established the nearest distinctions:

- [#1287](https://github.com/CrocSwap/integer-mult-bounds/pull/1287), head `8606df05eca3c65032361c74ee3dbc2630b0cdc4`: order-three cyclic output frames with a three-state recurrence.
- [#1199](https://github.com/CrocSwap/integer-mult-bounds/pull/1199), head `85ac2880ea5bda44f28632b71ffa2b4e35ec6b7d`: cancelling signed diagonal readers and checkpoint-correct partitions.
- [#1203](https://github.com/CrocSwap/integer-mult-bounds/pull/1203), head `38459854fd5bb93099871a84324413bd0e338d8e`: noncommuting reader restoration under a paid dirty wrapper.
- [#1227](https://github.com/CrocSwap/integer-mult-bounds/pull/1227), head `fea44a4d0d78570021d854881d15d9efe62e1713`: sharp output-sign restoration for a fixed-reader boundary.
- [#1246](https://github.com/CrocSwap/integer-mult-bounds/pull/1246), head `5ed1ce17a9df4c853a93f9758ba903d2620c3443`: paid nonzero reader-defect repair.
- [#1303](https://github.com/CrocSwap/integer-mult-bounds/pull/1303), head `bddf5ba853f1b798d021454362e53338c819f150`: exact ten-primitive lower bound for a restored local disjoint 2+3-cycle boundary, including finite dirty donors. It changes a local restoration word; this scope instead changes the child-facing frame contract.
- [#805](https://github.com/CrocSwap/integer-mult-bounds/pull/805), head `ef34bd5b32fcf2d5640dae90b5b8f28488738d5a`: determinant-balanced cyclic recurrence leaves with a different coefficient and precision contract.
- [#846](https://github.com/CrocSwap/integer-mult-bounds/pull/846), head `4bdb3e755ea7118170a7a34764df955880dec282`: stride-dependent cyclic inverse scheduling; different object and tape recurrence.
- [#1252](https://github.com/CrocSwap/integer-mult-bounds/pull/1252), head `1b9db2faf82b8866b0bb0164501c9a627c7dca9b`: the literal `Z4` match is an exact owner rank-pair phase polynomial for `48+48 -> 45+51`, not a four-cycle frame.

The targeted local-ledger search for these PR numbers returned no rows; the live heads and bodies above therefore control this overlap check. No selected construction or prior negative result is changed.
79 changes: 79 additions & 0 deletions research/z4-paired-frame-restoration-seed56/PREDECESSORS.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,79 @@
{
"schema_version": 1,
"new_work_commit_base": "bce54fb3f895cbe5cd4c0eb477c3afeed100190f",
"predecessors": [
{
"pr": 1287,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1287",
"head": "8606df05eca3c65032361c74ee3dbc2630b0cdc4",
"author": "Ross Richardson",
"contribution": "Order-three cyclic output frames, equivariant-reader admission, and a three-state boundary recurrence with a full dirty echo.",
"relationship": "Closest method predecessor. This package uses a paired order-four frame, different ADD/NEG phase costs, and a separate five-child invoice; no implementation is copied."
},
{
"pr": 1227,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1227",
"head": "fea44a4d0d78570021d854881d15d9efe62e1713",
"author": "Sagar Thakrar",
"contribution": "Paid output-sign frames and exact restoration costs for fixed signed-permutation readers.",
"relationship": "Retained credit for output-frame restoration; the paired order-four permutation and five-child recurrence are distinct."
},
{
"pr": 1199,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1199",
"head": "85ac2880ea5bda44f28632b71ffa2b4e35ec6b7d",
"author": "Sagar Thakrar (PR body credits Navinor / Aaron Wang)",
"contribution": "Checkpoint-correct coalescing of restoration for fixed signed diagonal readers.",
"relationship": "Related dirty-restoration baseline; its cut planner does not contain the paired order-four phase model."
},
{
"pr": 1203,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1203",
"head": "38459854fd5bb93099871a84324413bd0e338d8e",
"author": "mightyOil",
"contribution": "Noncommuting reader-restoration identities with a paid dirty-state wrapper.",
"relationship": "Related reader-interface work; this package uses fixed block selectors and explicitly prices cyclic frame gates."
},
{
"pr": 1246,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1246",
"head": "5ed1ce17a9df4c853a93f9758ba903d2620c3443",
"author": "Hanson Sun",
"contribution": "Paid repair of nonzero reader defects at canonical cuts.",
"relationship": "Related correction mechanism; not used in the paired-cycle frame construction."
},
{
"pr": 1303,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/1303",
"head": "bddf5ba853f1b798d021454362e53338c819f150",
"author": "mightyOil",
"contribution": "A scoped exact ten-primitive lower bound for a restored disjoint 2+3-cycle boundary, including arbitrary finite restored dirty donors.",
"relationship": "Relevant negative result for restored local cycles. It does not include this package's equivariant child maps or coupled P/Q frame phase schedule."
},
{
"pr": 805,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/805",
"head": "ef34bd5b32fcf2d5640dae90b5b8f28488738d5a",
"author": "Victor Liwentao",
"contribution": "A determinant-balanced cyclic recurrence leaf with coefficient, precision, and dirty restoration accounting.",
"relationship": "A different cyclic recurrence and data type; the present frame action is a finite signed-permutation interface."
},
{
"pr": 846,
"url": "https://github.com/CrocSwap/integer-mult-bounds/pull/846",
"head": "4bdb3e755ea7118170a7a34764df955880dec282",
"author": "Rohan S. Guliani",
"contribution": "Stride-independent cyclic inverse cost on sequential tapes.",
"relationship": "Different object and tape recurrence; cited to distinguish the word cyclic in the overlap review."
}
],
"selected_lineage": {
"selected_result": "certificates/selected-result.json",
"selected_supplier_pr": 1094,
"complex_word_predecessor_pr": 1087,
"bit_supplier_pr": 589,
"bit_lineage_pr": 555,
"relationship": "Source-pinned context only. The finite q4 interface is not integrated into the selected suppliers or profile."
},
"code_reuse": "No predecessor implementation or source archive is copied."
}
Loading
Loading