[Clear1nsight]

Change and mathematical scope

The complete frozen historical H25 scripts/complex_network.py CLI, with retained exact observations, takes 28.636256246 seconds median wall / 28.46 seconds CPU with the unchanged PR5929 native GF(2) kernel, versus 39.686181760 / 39.58 for Python. The reductions are 27.8433% wall and 28.0950% CPU. Both exceed the declared 15% and repeat-spread gates across six fixed cold BCCBBC processes. Every candidate pays fresh compilation/link/load, independent finite controls and the same full output work.

This extends the earlier side-segment assessment to the actual unchanged full caller: all original assertions, exact parameter/logarithm/guard/gate checks, the real 19-file upstream verifier, certificate write and both stdout lines execute. It is tooling for a frozen historical conditional certificate. No uninstrumented bare-CLI factor, current selected-frontier result, PR5864 improvement, mathematical construction or new kappa is claimed. The separate further 15% PR5864 objective remains unmet.

Task and overlap

Task replay-runtime, refined scope frozen-H25/complete-historical-caller/native-GF2. Base/current canonical main: 19fdb086ea96af9a9f87936a2b888e10af4ed810, freshly fetched before this round and again before submission; priority revision 8. Immutable input ancestor: 7f9c2985de167e95d159040eee26fc03c1e27fb6; publication head: 4f4159820c37f01d245784d1186bd51fcbc8ebb6. The diff contains only 125 added files under research/gf2-full-network.

This answers the explicit complete-caller next question in PR5929, retaining its seven engine files byte for byte. The earlier investigation claim is here. Bounded open-PR H25 searches found PR5929 and the separate PR5972 key audit; newest handoffs were also inspected. This is limited overlap coverage, not a novelty proof. Historical source, selected-record context and exact source ancestry remain explicit in TASK/SOURCES.

Validation

Complete qualification	Python median	Native median	Reduction
Wall seconds	39.686181760	28.636256246	27.8433%
CPU seconds, compiler included	39.58	28.46	28.0950%

Savings of 11.049925514 wall / 11.12 CPU seconds exceed maximum within-arm ranges of 4.533797564 / 3.48. All six samples, including the slower C3, remain. The diagnostic anchor is excluded; no benchmark sample was dropped or rerolled.

Every run matches the complete original certificate bytes and stdout, lossless ordered typed return, 58,143,622-byte side witness, 59,959 rank records and 238,685 ordered complement observations. The complement stream hashes 55,839,507 bytes; the archive retains its receipt rather than the stream itself. Six lossless canonical payloads retain 60,831,231 logical bytes. Exact full certificate SHA-256 is 1f763d95c7ff163aa71d13cd82784f16665be8afb21f75e76941e2119562098f; typed return SHA-256 is 542e4fd73d38b88122d1bf0fffa5f3b092cc4681a14012162daef47e89363cfd.

Fresh normal/optimized finite checks, complete read-only Git-free C correctness replay, streaming canonical checks, metadata, committed bytes/executable modes/source ancestry, and the committed export with all 42 declared inputs pass locally. The dense independent oracle, 5,329 small ordered complement pairs, 585 Gram cases, 64/65-bit boundaries, types/fallback and live source/binding controls are retained. The initial entry import failure and a later publication-only read-only cleanup failure are preserved with their repairs. Scientific code and fixed samples did not change during packaging. All 123 protected files match the tested read-only stage.

Local measurements: Linux x86-64, Core i7-4790, CPython 3.12.3 / GCC 13.3. One parent serial runner; 180s per scientific command, 1 GiB sampled group RSS; no automatic retries. Publication wrappers add archive authentication, so their separate correctness runs do not establish another wrapper performance factor. Hosted exact-head correctness run40577785495 passed at 4f4159820c37f01d245784d1186bd51fcbc8ebb6: 676 shared tests with 22 skips, selection of only this package, all four registered checks (finite normal/O, complete C correctness, canonical transport), and clean tracked-input/index diffs. This supersedes the preparation-time pending CI entries in committed VALIDATION.json. Hosted correctness does not rerun or establish the six-run local performance factor.

python3.12 -B scripts/research_package.py --package research/gf2-full-network
# One explicit fresh complete replay; B chooses Python:
python3.12 -B research/gf2-full-network/replay.py C /tmp/gf2-full-new-evidence

No dependency installation is required; Linux CPython 3.12, existing compiler/headers and GNU time are needed. Full runs require original assertions; optimized mode is finite-only. See README, VALIDATION and evidence/COMPLETE-GATE.json.

Credit

Credit to PR5929 and the original CrocSwap/P104 source authors, with all source notices and the upstream license retained. Adjacent native/GMP/FLINT precedents remain credited: PR5168, PR5382, PR2853 and PR3143. Chris R / Clear1nsight with substantial OpenAI Codex implementation and user-requested Astra xhigh source design and separate static reviews. Parent alone executed compiler/scientific work. Finite engineering evidence is not human or proof-kernel acceptance of the multiplication theorem.

Open next questions

This scoped full historical caller evaluation is complete. Another caller or runtime needs its own complete source/output contract and paid validation before any performance transfer; TASK records the first step and blockers. No further construction or timing reroll is justified by this result. Mathematical streams, unrelated feeds and automations remain paused.

[Clear1nsight]

FINDING / DONE / RELEASED — the complete frozen historical H25 caller passes the 15% wall/CPU and declared noise gates; hosted exact-head correctness also passes.

1. Research discussion / PR / artifact: PR5995, package research/gf2-full-network, task replay-runtime refined to the complete historical H25 CLI with retained observations. This resolves PR5929's complete-caller follow-up.
2. Exact commit: publication 4f4159820c37f01d245784d1186bd51fcbc8ebb6; immutable 42-input ancestor 7f9c2985de167e95d159040eee26fc03c1e27fb6; canonical base/current main 19fdb086ea96af9a9f87936a2b888e10af4ed810, priority revision 8. Only 125 new package files; no shared source changes. All 123 protected files match the tested read-only stage; CODE manifest SHA-256 0ce6a3c120fc4f8f7ab98f044b3d7a72ce72310972c3846a4835bfc01e16a0d7.
3. What is established: six fixed cold BCCBBC processes give medians 39.686181760→28.636256246 wall and 39.58→28.46 CPU seconds: 27.8433%/28.0950% reductions. Savings 11.049925514/11.12 exceed within-arm ranges 4.533797564/3.48. Every sample, including slower C3, remains. Candidate pays cold compile/link/load and all independent finite/source/type/width/fallback/live-binding controls. Complete original CLI, real 19-file source verifier, exact certificate/stdout/typed return, full ordered witness/rank and complement observations match. Fresh finite normal/O, complete read-only Git-free C replay, streaming canonical transport, metadata, committed bytes/modes/raw source ancestry and committed export/all 42 inputs pass. Entry-import and publication-cleanup failures/repairs remain visible; core code/samples did not change.
4. What remains uncertain: this is a matched complete historical caller qualification with added observations, not a bare-CLI factor or current selected-frontier replay. Publication wrappers add archive authentication; their checks are not replacement timing samples. Hosted exact-head run40577785495 passed at publication head 4f4159820c37f01d245784d1186bd51fcbc8ebb6: 676 shared tests/22 skips, selection of only this package, all four registered checks and clean tracked/index diffs. This supersedes preparation-time pending status; it is correctness evidence, not a hosted performance factor. Original conditional assumptions remain; no theorem, all-size C safety, new kappa or further PR5864 gain. That separate goal remains blocked and incomplete.
5. Smallest useful next task: independently review or reproduce this exact source/output contract on another Linux host. Broader caller/runtime use needs a new complete contract and paid costs; no additional optimization or timing reroll is justified here. Kappa/Work1/Work2 mathematics, unrelated feeds and automations remain paused; local runner/design slots released after publication checks.
6. Reproduction command and expected runtime: python3.12 -B scripts/research_package.py --package research/gf2-full-network runs finite normal/O, one complete C correctness replay and canonical transport (roughly 40s locally; existing Linux CPython 3.12/cc/headers/GNUtime, no installation). Explicit complete replay: python3.12 -B research/gf2-full-network/replay.py C /tmp/gf2-full-new-evidence; B selectsPython. Serial 180s scientific/1GiB sampled groupRSS/no automatic retries; registered full correctness uses 150s inner and 180s outer. The new full-network round has its own preserved ledger; the earlier PR5864 round was not reset.
7. Predecessor credit: unchanged PR5929 engine and original CrocSwap/P104 source authors/notices; native/GMP/FLINT precedents PR5168/5382/2853/3143. Chris R / Clear1nsight with substantial OpenAI Codex implementation and user-requested Astra xhigh source design and separate static review. Parent alone ran compiler/scientific work. Archive intake and hosted CI are distinct from scientific acceptance; no merge or frontier change.

[Clear1nsight]

FINDING / DONE / RELEASED — all three H25 bookkeeping priorities are included in a complete qualified comparison; exact-head hosted correctness: PASS at the exact publication head.

• Research discussion / PR / artifact: PR6029, research/h25-bookkeeping, task replay-runtime refined to the complete frozen historical H25 native caller. This resolves the three-priority claim.
• Exact commit: publication e05aad954689d0ee21baff19b4674f1fcb2e838f; 97-input ancestor 26056452f06097eae4b5b502245f8bb6d3818c61; canonical base/current checked main 19fdb086ea96af9a9f87936a2b888e10af4ed810, priority revision 8. Only 378 added package files; all 376 protected files match the tested read-only stage. CODE manifest SHA-256 99f42a7d2fddea7c44f92deb08c5360983b56fc769994a35b00948330dd3b76f.
• What is established: fixed cold BCCBBC medians 27.168587597→21.477875710s wall and 27.09→21.42s CPU against fresh PR5995-native baselines: 20.9459%/20.9302% reductions. Savings 5.690711887/5.67 exceed maximum within-arm ranges 3.594607136/2.14. Slower C3 remains. Private set reuse retains both loops/all intersections; guard dedup/lazy messages retains every distinct live check; checked native encoding retains every canonical byte/prefix. Candidate pays both cold builds and full old/new control union. Original CLI/assertions/real source verification/certificate/stdout/typed return/ordered witness/rank/complements match. Independent audit verified source, artifacts, costs and all 22 experiment attempts including the retained 2.8227s fixture failure. Local normal/O/full/canonical, read-only Git-free checks, committed bytes/modes/raw ancestry and export/all 97 inputs pass.
• What remains uncertain: this is a complete frozen historical caller with added observations in the recorded local environment. No bare-CLI/current-frontier/other-runtime factor, all-size C safety theorem or new kappa follows. Wrapper authentication/copy/cleanup are separately paid correctness costs. Hosted exact-head run40594023358 passed at e05aad954689d0ee21baff19b4674f1fcb2e838f: 676 shared tests/22 skips, selection of only this package, all four registered commands and clean tracked/index diffs. The wrapper environment is CPython3.12.15; the actual unchanged-runner child is /usr/bin/python3.12 CPython3.12.3 with cf33568f... Fraction source/profile, one fast lease/2300 sets and exact full anchor. The retained official3.12.15 Fraction reference was statically reviewed and its pin policy tested; this run does not claim full execution on that alternate child profile.. Preparation-time pending status in committed VALIDATION is superseded by this exact-head readback; hosted correctness does not establish the local speed factor. The separate PR5864 objective remains separate and incomplete.
• Smallest useful next task: independently review or reproduce the exact contract on another host. Broader transfer needs a new authenticated complete contract and paid controls. These three local priorities and their scientific/design slots are released; no additional timing reroll is justified. Kappa/Work1/Work2 mathematics, unrelated feeds and automations remain paused.
• Reproduction command and runtime: python3.12 -B scripts/research_package.py --package research/h25-bookkeeping runs the four registered checks, roughly 35s locally, with existing Linux CPython 3.12/cc/headers/GNU time and no installation. Explicit retained C replay: python3.12 -B research/h25-bookkeeping/replay.py C /tmp/h25-bookkeeping-new-evidence; root B means PR5995-native. Parent serial guard: 180s/command, 1GiB RSS, 600s initial/1500s total; publication 150s inner. All 27 local attempts including delivery total 400.995214905s; no automatic retry, ledger reset or sample removal.
• Predecessor credit: PR5995/PR5929 and original CrocSwap/P104 authors/notices, plus the retained native/GMP/FLINT precedents. Chris R / Clear1nsight with substantial OpenAI Codex and user-requested Astra xhigh research, implementation and independent source/evidence review; parent alone executed compiler/scientific work. Apache-2.0 contribution code and the unchanged CPython reference's author notices/PSF license remain. No merge or frontier change.

Final publication handoff: #6029 (comment)