Integration: admit PR883 X4 primitive lowering through selected invoice and assembly (no new kappa)

research/x4-primitive-admission/CLAIM.md

@@ -0,0 +1,17 @@
+# User-selected PR883 integration
+
+The user explicitly selected follow-ups to PR883, PR890, PR853, PR882 and PR889.
+This is the first contribution in that goal. No random track or new topology
+is inferred. The local object is PR883's 56 zero-frame target-pair expansions;
+the method is concrete primitive emission, separate exact block verification,
+source-bound native/phase/wrapper/owner transport, complete invoice and assembly.
+
+Stopping rule: freshly reproduce the original package, verify the full selected
+native/global baseline with complete ancestry coverage, pass explicit lowering
+and complete accounting, reject prefix/invoice corruptions, check relocation,
+then submit with preserved sources/credit and no unilateral selected change.
+
+This adds integration and accounting evidence, not a priority claim for the
+factorization, pair optimum or compiler primitive alphabet. Concurrent PR901
+changes the later compensation and is neither imported nor added to this bill.
+Main remains pinned at f567884a; the final PR883 head is verified before publication.

research/x4-primitive-admission/EXPECTED.json

@@ -0,0 +1,10 @@
+{
+  "status": "PASS_X4_PRIMITIVE_GLOBAL_ADMISSION",
+  "base": "f567884aeb8851b6e794af1184119a3bfbe1e631",
+  "predecessor": "b31a870a3842b37f6ab97767c506ce91783632a2",
+  "kappa": "1661634653669567/1000000000000000000",
+  "saved_selected_real_component_primitives": 56448,
+  "saved_coefficient_work_operations": 36288,
+  "global_prefix_bits": 257,
+  "physical": 7682
+}

research/x4-primitive-admission/FRESH-REPLAY.json

@@ -0,0 +1,190 @@
+{
+  "status": "PASS_FRESH_SELECTED_REPLAY_AND_PRIMITIVE_REFINEMENT",
+  "baseline_binding": {
+    "status": "PASS_REBUILT_PR589_PR799_COMPOSITION",
+    "kappa": "1661634653669567/1000000000000000000",
+    "decimal": "0.001661634653669567",
+    "ancestry": {
+      "status": "PASS_REBUILT_PR589_PR741_COMPOSITION",
+      "kappa": "1661202643272239/1000000000000000000",
+      "decimal": "0.001661202643272239",
+      "bit_input_sha256": "b9f0e5ca038abe837b0007c8fe006f366a6f02009244ac57705a8a200e2301f4",
+      "bit_result_sha256": "28795393651ef4fbbc6232eb2b32f7b39b91fbd5ea6d2315583f350714698055",
+      "complex_receipt_sha256": "68600c734560bb155037e3973f7fbfb2fd1cb678485f8ac128c0b1e1bb1840a0",
+      "complex_assembly_sha256": "35ccaac474236b2e244f02cc599fbe4ddee783713f37ea623842ecb8de31f209",
+      "scope": "Rebuilt finite suppliers and exact composition. All-size/compiler, common-label/mutable-source, whole-call hosting, precision/tape and analytic transfer hypotheses remain conditional; not a full Lean theorem."
+    },
+    "package_manifest_sha256": "e1b7998c8f7cb24f8674d250d0510f2b5b351b2ed0b35f17a17e6bacf3d9824b",
+    "assembly_sha256": "a7ab72bc4a3a00bd2147ae30dcf67448d0dfd69ab57e8f27f5ef520d0a560d01",
+    "independent_sha256": "932930a48de3a0d2b68fb3d1272521b6a6692b7ead886dbda608823f49030d35",
+    "price_sha256": "04bff9778b12a8dd878e0bf304c163d93aea09f3b7b3ebc878dab66e23054dd9",
+    "scope": "Exact finite X4 scalar/frame/phase/owner/invoice/assembly checks, bound to rebuilt PR589 and SW4 ancestry. Inherited all-size/compiler, common-label/mutable-source, whole-call hosting, precision/tape and analytic assumptions remain; not a full Lean theorem.",
+    "reused_ancestry": true
+  },
+  "guarded_public_result": {
+    "status": "PASS_COMPLETE_PR799_REPLAY_COVERAGE",
+    "kappa": "1661634653669567/1000000000000000000",
+    "decimal": "0.001661634653669567",
+    "coverage_contract": "pr741-sw4-v1",
+    "contract_sha256": "7acd5033bc78e3f473718269edaa267c2785d84fbe95ccf6815a33dcf58c2756",
+    "selected_binding": {
+      "status": "PASS_REBUILT_PR589_PR799_COMPOSITION",
+      "kappa": "1661634653669567/1000000000000000000",
+      "decimal": "0.001661634653669567",
+      "ancestry": {
+        "status": "PASS_REBUILT_PR589_PR741_COMPOSITION",
+        "kappa": "1661202643272239/1000000000000000000",
+        "decimal": "0.001661202643272239",
+        "bit_input_sha256": "b9f0e5ca038abe837b0007c8fe006f366a6f02009244ac57705a8a200e2301f4",
+        "bit_result_sha256": "28795393651ef4fbbc6232eb2b32f7b39b91fbd5ea6d2315583f350714698055",
+        "complex_receipt_sha256": "68600c734560bb155037e3973f7fbfb2fd1cb678485f8ac128c0b1e1bb1840a0",
+        "complex_assembly_sha256": "35ccaac474236b2e244f02cc599fbe4ddee783713f37ea623842ecb8de31f209",
+        "scope": "Rebuilt finite suppliers and exact composition. All-size/compiler, common-label/mutable-source, whole-call hosting, precision/tape and analytic transfer hypotheses remain conditional; not a full Lean theorem."
+      },
+      "package_manifest_sha256": "e1b7998c8f7cb24f8674d250d0510f2b5b351b2ed0b35f17a17e6bacf3d9824b",
+      "assembly_sha256": "a7ab72bc4a3a00bd2147ae30dcf67448d0dfd69ab57e8f27f5ef520d0a560d01",
+      "independent_sha256": "932930a48de3a0d2b68fb3d1272521b6a6692b7ead886dbda608823f49030d35",
+      "price_sha256": "04bff9778b12a8dd878e0bf304c163d93aea09f3b7b3ebc878dab66e23054dd9",
+      "scope": "Exact finite X4 scalar/frame/phase/owner/invoice/assembly checks, bound to rebuilt PR589 and SW4 ancestry. Inherited all-size/compiler, common-label/mutable-source, whole-call hosting, precision/tape and analytic assumptions remain; not a full Lean theorem."
+    },
+    "reused_ancestry": true,
+    "scope": "Required evidence presence, ordered stages and existing hash/mathematical bindings. Not execution authentication or a full multiplication proof."
+  },
+  "refinement": {
+    "status": "PASS_X4_PRIMITIVE_GLOBAL_ADMISSION",
+    "base": "f567884aeb8851b6e794af1184119a3bfbe1e631",
+    "predecessor": "b31a870a3842b37f6ab97767c506ce91783632a2",
+    "kappa": "1661634653669567/1000000000000000000",
+    "saved_selected_real_component_primitives": 56448,
+    "saved_coefficient_work_operations": 36288,
+    "global_prefix_bits": 257,
+    "physical": 7682,
+    "lowering_sha256": "565fe81544f6d7e58bb17c12f37beae03a9397f66b0eae77e1ee6a4ee81743c9",
+    "invoice_sha256": "2751c2fd4e84aa90ecba0bfc30bc90d68f4b6a30e08b138df7834dce4613f4f1",
+    "assembly_sha256": "1372e5067c015b122b05d5dac551df96b66a923d81b7639b8305ceb73ba528b1",
+    "scope": "Fresh selected native/phase/wrapper/owner admission plus certified explicit primitive refinement and complete invoice/assembly. No new kappa; inherited all-size and analytic assumptions remain."
+  },
+  "native_steps": [
+    {
+      "step": "jx",
+      "exit_code": 0,
+      "seconds": 3.594
+    },
+    {
+      "step": "jx-control-C0_identity",
+      "exit_code": 0,
+      "seconds": 3.172
+    },
+    {
+      "step": "jx-control-C1_741_only",
+      "exit_code": 0,
+      "seconds": 2.562
+    },
+    {
+      "step": "jx-control-M1_one_pair_448",
+      "exit_code": 0,
+      "seconds": 3.047
+    },
+    {
+      "step": "jx-control-N1_drop_preread",
+      "exit_code": 0,
+      "seconds": 2.0
+    },
+    {
+      "step": "jx-control-N2_wrong_sign_preread",
+      "exit_code": 0,
+      "seconds": 3.718
+    },
+    {
+      "step": "jx-control-N3_plain_join_moved_reads",
+      "exit_code": 0,
+      "seconds": 3.656
+    },
+    {
+      "step": "jx-control-N4_raise_dim7_at_855",
+      "exit_code": 0,
+      "seconds": 2.563
+    },
+    {
+      "step": "layer-end-controls",
+      "exit_code": 0,
+      "seconds": 1201.297
+    },
+    {
+      "step": "native",
+      "exit_code": 0,
+      "seconds": 13.5
+    },
+    {
+      "step": "ledgers",
+      "exit_code": 0,
+      "seconds": 9.266
+    },
+    {
+      "step": "local-replay",
+      "exit_code": 0,
+      "seconds": 80.735
+    },
+    {
+      "step": "wrappers",
+      "exit_code": 0,
+      "seconds": 33.422
+    },
+    {
+      "step": "phases",
+      "exit_code": 0,
+      "seconds": 2041.469
+    }
+  ],
+  "global_steps": [
+    {
+      "name": "restore-parent",
+      "seconds": 10.266,
+      "exit_code": 0,
+      "log_sha256": "3aea2a4b68e7aaeadd9ffef649be0d25e2dff1535d9a61bffeaf6d92e55ba6a8"
+    },
+    {
+      "name": "bind",
+      "seconds": 5.859,
+      "exit_code": 0,
+      "log_sha256": "78a081673d6b4fde10a86e6f767d53648b868d1f0cac2441d12951434ad7341b"
+    },
+    {
+      "name": "emit",
+      "seconds": 31.0,
+      "exit_code": 0,
+      "log_sha256": "e2e13b5dd252e7209473f068993c91615705ba63614de998e233909ffa8f8b7d"
+    },
+    {
+      "name": "audit",
+      "seconds": 116.906,
+      "exit_code": 0,
+      "log_sha256": "7f392b1834727b565c6f97e8c8cfba61e439db0742797d217913cfdb55f16117"
+    },
+    {
+      "name": "invoice",
+      "seconds": 15.468,
+      "exit_code": 0,
+      "log_sha256": "06023deeec03cb062e55fc5c8fe2ad59efbfccd0a9af8836cf1356a0d16157dc"
+    },
+    {
+      "name": "assembly",
+      "seconds": 2.282,
+      "exit_code": 0,
+      "log_sha256": "ab65a6031607ba15c7f9c8a70f2bc5a75f02492e7515eaa2a8b23e231c2fb6d3"
+    },
+    {
+      "name": "admission-controls",
+      "seconds": 2.985,
+      "exit_code": 0,
+      "log_sha256": "34527ddfe1e53bec71dc5759eb9a26417be6d4e93e76cbf80575d9bafebd876b"
+    },
+    {
+      "name": "bit-supplier-controls",
+      "seconds": 4.422,
+      "exit_code": 0,
+      "log_sha256": "c648542cd6dddbb154af7d02a41aaed1025603d9a4ab54f7c249d2b683e22249"
+    }
+  ],
+  "scope": "Fresh PR799 native/global replay with complete explicitly verified PR741 ancestry reuse; not a fresh bit or SW4 reconstruction."
+}

research/x4-primitive-admission/PROOF.md

@@ -0,0 +1,67 @@
+# Compiler refinement and admission argument
+
+This does not replace the recursive supplier or its logical word. It replaces
+the primitive expansion of 56 disjoint, frame-zero target blocks. The literal
+forward expansion has matrix `((-1/2,1/2),(1/2,1/2))`; the backward expansion
+has matrix `((-1,1),(1,1))`, the inverse transpose. The independently implemented
+two-row checker computes each emitted matrix from the instructions themselves.
+It rejects a nonzero frame, a register outside the pair, a private scratch read,
+an incorrect coefficient type, or an unequal endpoint.
+
+Both targets begin at the zero frame. Every primitive acts on those same targets
+at that frame, with no recursive query or new carrier. At the end of each block,
+all native registers therefore have exactly their retained values. The original
+coefficient work can retain its entering dirt across the fused interval; the
+unchanged suffix explicitly clears it before any read. Complete emitted scalar
+replay checks that property with nine independent dirty private columns in
+addition to all 728 native columns. Exact clearing is not an initialization
+assumption.
+
+Each physical non-scalar event is retained literally and in order. The complete
+source-bound native/frame/phase/wrapper receipts and controller audit are required
+before refinement. Equality at each block boundary transports those operator
+obligations to the primitive implementation: the changed block has no phase or
+frame transition to readmit. This is a certified compiler refinement, rather
+than a claim that the historical verifier natively parses the new primitive
+format. That verifier checks the logical interface; this checker verifies the
+concrete implementation and their relation.
+
+Original scalar bounds charge eleven primitives per component per pair. The
+new literal expansions charge four. For 56 pairs and two real components this
+is `56*2*(11-4)=784` per orientation. The original two ADD expansions use nine
+coefficient-work operations per component per pair; the new block uses none.
+That is 504 fewer operations in the predecessor's coefficient-work convention.
+The original invoice has 72 complete calls, so its selected sums fall by 56,448
+and 36,288 respectively. Coefficient work is already included in the primitive
+sum; the reductions are not added together.
+
+Logical macro, adapter, center, wrapper, child, stock, dispatch, owner, source
+access, routing and old pervertex overcharge reserves are retained. The compiler
+can expand a logical three-macro block jointly using the certified four-operation
+word. All added primitive work is explicitly counted, and no new physical
+frame adapter is required inside a frame-zero target block. Existing adapter
+reserves cover the unchanged boundary/physical event interface.
+
+Complete emitted exact row-L1 and fractional-grid maxima agree with PR883's
+retained whole-program comparison. The isolated forward pair temporarily grows
+from 1 to 2, but the complete original absolute-envelope reserve already covers
+the new exact maxima. Both final native values and the old logical prefix
+contract remain valid at every block boundary. The refinement retains that
+larger reserve and checks exact intermediate magnitude and denominator bounds;
+it does not infer a rounded physical execution theorem from rational replay.
+
+The retained invoice function independently validates the original fresh
+provider before its counts can be refined. The new lowering receipt and concrete
+program digest are part of the generated provider contract. The full invoice
+function then recomputes every selected call, parent/pervertex allowance, stock,
+prefix, router, literal, mass, precision, tape, row and halving guard. A separate
+candidate SPEC changes only the refined provider contract and local-count fields.
+The unchanged assembly executes its conservative admission, exact moment
+engines, bit surcharge/deficit checks, bootstrap, native correction and 47 strict
+outer constraints. The recursive profile and final kappa are unchanged.
+
+All original source bytes remain immutable. PR865's source-versioned coverage
+gate additionally rejects missing required evidence before the original binder.
+These are finite operator/compiler/accounting checks under inherited all-size,
+realization, hosting, precision/tape and analytic hypotheses, not a complete
+integer multiplication theorem.

research/x4-primitive-admission/README.md

@@ -0,0 +1,77 @@
+# Admit PR883's explicit X4 primitive lowering
+
+Finite compiler refinement and complete accounting; no new kappa.
+
+Philipp's PR883 gives four-primitive implementations of the selected X4 word's
+56 zero-frame precondition pairs. This package emits those literal instructions
+inside a complete primitive program, preserves every physical frame, phase,
+center and wrapper event, and binds the refinement to a complete selected replay.
+An independent two-row matrix check validates every emitted pair. Complete
+737-column scalar replay checks the emitted streams and all private clearing.
+
+The original macro ledger remains the logical proof interface. Its native
+frame, phase, wrapper and whole-owner admission is valid because every changed
+pair has the identical endpoint at frame zero, uses only the same two targets,
+and introduces no recursive child or scratch register. All logical macros and
+compiler adapter reserves remain charged. The new concrete implementation is
+explicitly distinguished from the original macro ledger; see PROOF.md.
+
+The complete 72-call invoice saves **56,448 real-component primitives** and
+**36,288 coefficient-work operations**. The latter uses the predecessor's
+per-component accounting convention, not an additional saving to add to the
+primitive bill. Physical stock remains 7682; the conservative prefix remains
+257 of 512 bits. All finite guards and the unchanged exact assembly pass at
+`1661634653669567/1000000000000000000`. No runtime measurement is claimed.
+
+## Reproduce
+
+Python 3.11+, standard library only for the lowering/admission package. A full
+bit reconstruction additionally needs the predecessor's C++17/Boost toolchain.
+Supply a complete selected PR799 replay and its actual PR741 ancestry directory:
+
+```sh
+python -I -B research/x4-primitive-admission/admit.py \
+  --replay /completed/frontier799/replay --ancestor /completed/frontier741 \
+  --out /fresh/external/refined-admission
+python -I -B research/x4-primitive-admission/test_controls.py \
+  /fresh/external/refined-admission
+```
+
+If using the original public PR799 worker rather than PR865's guarded runner,
+pass its work directory directly. The unchanged vendored PR865 coverage gate
+checks all required ancestry stages/artifacts and selected-stage logs/results
+before the mathematical binder. An existing or in-repository output is refused.
+Native admission intentionally rejects disabled assertions.
+
+`lowering/PRIMITIVE-PROGRAMS.json.gz` is the emitted concrete implementation;
+`LOWERING.json` retains its exact endpoint, prefix, paid-profile and source
+bindings. `invoice/` holds the complete recomputation. `candidate-package/`
+contains a separate copied assembly context and generated refinement contract.
+`ASSEMBLY.json` and `RESULT.json` are published only after assembly succeeds.
+Original source packages, selected records and defaults are never edited.
+
+Six emitted-prefix corruptions per orientation, immutable source drift and three
+assembly controls must fail. The retained PR883 controls also reject all 672
+pair deletion/sign mutants. The original source package's independent pair
+proof and seven source/interface controls were freshly reproduced separately.
+
+## Credit and scope
+
+PR883 at `b31a870a3842b37f6ab97767c506ce91783632a2` is by Philipp (@mightyOil).
+Its exact original package is retained in `vendor/pr883`; no priority for the
+factorization, gatewise comparison or optimality proof is claimed. PR865 supplies
+the required-evidence gate, crediting Lucas (@ljste)'s PR838 and concurrent PR847.
+Selected native/phase/controller/invoice/assembly sources retain Jeremy Kothe,
+Chafik Boukhalfa, icekylinx, DreamingOfClouds, Rohan Arun, eumemic and predecessor
+attribution. SOURCES.json pins the consumed sources and contexts.
+
+Overlap refresh through PR902 found ToikY's concurrent PR901, which changes
+the following arbitrary-helper compensation and preserves this precondition.
+Its candidate is not consumed; no combined saving or integration priority is
+claimed. Main remains at the pinned f567884a context.
+
+Prepared for **nightangelflowerwin-ops** with substantial **OpenAI Codex
+assistance** in integration, verification, mathematics and writing, Apache-2.0.
+No independent human review, new exponent, measured speedup or full multiplication
+theorem is claimed. All-size compilation, physical realization/hosting,
+precision/tape and analytic transfer remain inherited assumptions.