Research lane D's family-gate cell and batch (POW-02, ROT-07, ROT-08 partial; recorded by the steward at the lane's ask)

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs 2026-10-08 19:35:58 +00:00
parent 6b0bc3aee4
commit 23cac34c77
4 changed files with 169 additions and 35 deletions

View file

@ -283,6 +283,16 @@ Rule: a case maps to a cell only where the cell's tests visibly answer it; cover
- Cases:
- POW-07 Constrain any mixed-resource or FP32 branch: the branch ran and its claim failed: the FP32 branch costs the card more energy and widens the chip edge; the master's register row 12 reads FAIL, published; the branch is excluded and unreachable on master (the grep evidence stays)
### harness:family-gate
- Command: `IGNEUM_FAMILY_GATE=1 [IGNEUM_FG_CLASS=<class string>] igneum_pow_test --ignored family_gate_era_census (the class v5 crate plus docs/analysis/class-v6/logs/harness-*.diff; one TSV row per drawn era); attack-f8 census --seeds 64 --nonces 1048576 --control window --by-site against the same crate`
- Box class: release (the family gate's top class on every box)
- Fixtures: F0, F1
- Cases:
- POW-02 Validate generated programs and index folding: partial: the acceptance rule's measured rates on drawn classes (the per-candidate rejection, the attempt histogram, the exhaustion count against the independent-attempt arithmetic, the (c'') and (c''') floors' refuse rates and the pre-floor ratio spread per width), the per-site largest-bucket and index-bit statistics at the rule's own 2^20 sample with the seven known-failed seeds (p4, p8, p10, p34, p212, p225, p15), and the F8-form census at 2^20 on the F8 seed set where the attack-f8 mirror validates (0 of 96 warps mismatching); the semantics, termination and memory-bound half and the independent re-implementation are the suite's and the hash lane's; the effect-size limits are the family-gate.md thresholds, predeclared in its section 3
- ROT-07 Ablate redundant rotation layers: partial: the acceptance-side ablation rows (layer 8 off against the same-crate windowed control: the draw cost, the floors' costs, the hot-set and 1.2x ceiling, the bit-R read; the mixer and shape axes' per-era invisibility and verifier rows), docs/analysis/class-v6/family-gate.md sections 6.4 to 6.12; the adversarial cost and the GPU setup rows per layer are the hash lane's and the adversary lane's
- ROT-08 Pass the no-new-rules counterfactual: partial: the no-rescue control document (38,000 drawn eras of the parameter family through the acceptance, the bound arithmetic on the counts, the honest line of what the gate cannot see), family-gate.md sections 1 to 6; the competitiveness envelope and the economic half are ECO's
## Automated cases with no harness in the matrix (NOT RUN, the reason)
- GOV-02 Approve thresholds before results: the approval is recorded in the registry's approval field; the automated half (thresholds frozen before any run_status) is the gate rule landing by 21:00
@ -297,8 +307,6 @@ Rule: a case maps to a cell only where the cell's tests visibly answer it; cover
- ADV-07 Evaluate lifetime without forced obsolescence: lifetime rows are the adversary lanes' models
- ROT-03 Test miner-voted bring-forward governance: miner-voted bring-forward needs a vote harness on the fault network F4
- ROT-04 Resist seed selection and faster evaluators: seed-selection resistance is the census harness (the class v6 invention lane), not yet in the matrix
- ROT-07 Ablate redundant rotation layers: layer ablation is a research harness, not a release suite
- ROT-08 Pass the no-new-rules counterfactual: the no-new-rules counterfactual is a research harness
- ECO-02 Separate existing-owner and new-entrant viability: F7
- ECO-03 Let the specialist keep its sunk development: F7
- ECO-04 Model entry, exit and difficulty response: F7
@ -452,4 +460,4 @@ Rule: a case maps to a cell only where the cell's tests visibly answer it; cover
## Count
171 automated cases: 57 mapped to a cell, 164 NOT RUN with a reason.
171 automated cases: 59 mapped to a cell, 162 NOT RUN with a reason.

View file

@ -1359,31 +1359,30 @@
"manual_page": 24,
"owner_lane": "hash lane (a690540514aa453d7)",
"run_status": "NOT RUN",
"evidence_path": "build-1:/srv/artefacts/tas/201-7cfa422a-aa0e0f45/miner-9c844503/box2-pow.log",
"run_id": "201-7cfa422a-aa0e0f45",
"updated": "2026-10-08T19:32:50.856Z",
"evidence_path": "build-1:/srv/artefacts/tas/201-7cfa422a-aa0e0f45/miner-9c844503/box2-pow.log; docs/analysis/class-v6/family-gate.md",
"run_id": "family-gate-20261008",
"updated": "2026-10-08T19:35:58.408Z",
"evidence_record": {
"cell": "suite:pow",
"manifest_sha": "7cfa422a",
"coverage": {
"POW-02": "partial: program derivation, ds55 geometry, the v6 fold, the packs and the spec read-back tests; the independent re-implementation and the index-fold census are the hash lane's harnesses",
"POW-08": "partial: the freeze list and the generator version pin; the public-claim scrub is the site gate's",
"POW-07": "the mixed FP32 branch is excluded and unreachable: no class flag reaches an FP op in igneum-pow on master (the mixedfp lane's branch is not on master); evidence the grep of the emitters on master, recorded by the hash lane"
},
"at": "2026-10-08T19:32:50.856Z",
"method": "native",
"requirement_id": "POW-02",
"decision": "NOT RUN",
"reviewer": "",
"claim_impact": "",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the acceptance rule's measured rates on drawn classes (the per-candidate rejection, the attempt histogram, the exhaustion count against the independent-attempt arithmetic, the (c'') and (c''') floors' refuse rates and the pre-floor ratio spread per width), the per-site largest-bucket and index-bit statistics at the rule's own 2^20 sample with the seven known-failed seeds (p4, p8, p10, p34, p212, p225, p15), and the F8-form census at 2^20 on the F8 seed set where the attack-f8 mirror validates (0 of 96 warps mismatching); the semantics, termination and memory-bound half and the independent re-implementation are the suite's and the hash lane's; the effect-size limits are the family-gate.md thresholds, predeclared in its section 3",
"release_identity": {
"commit": "7cfa422a",
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
}
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
},
"in_progress_since": "2026-10-08T19:32:50.856Z",
"approvals": {
@ -1418,6 +1417,28 @@
"run_id": "201-7cfa422a-aa0e0f45",
"evidence": "build-1:/srv/artefacts/tas/201-7cfa422a-aa0e0f45/miner-9c844503/box2-pow.log",
"in_progress": true
},
"harness:family-gate": {
"requirement_id": "POW-02",
"decision": "NOT RUN",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the acceptance rule's measured rates on drawn classes (the per-candidate rejection, the attempt histogram, the exhaustion count against the independent-attempt arithmetic, the (c'') and (c''') floors' refuse rates and the pre-floor ratio spread per width), the per-site largest-bucket and index-bit statistics at the rule's own 2^20 sample with the seven known-failed seeds (p4, p8, p10, p34, p212, p225, p15), and the F8-form census at 2^20 on the F8 seed set where the attack-f8 mirror validates (0 of 96 warps mismatching); the semantics, termination and memory-bound half and the independent re-implementation are the suite's and the hash lane's; the effect-size limits are the family-gate.md thresholds, predeclared in its section 3",
"release_identity": {
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
}
}
},
@ -3074,11 +3095,29 @@
"owner_lane": "lane D family gate (a07a99a3788566af2)",
"run_status": "NOT RUN",
"evidence_path": "docs/analysis/class-v6/family-gate.md",
"run_id": "team-2026-10-08",
"updated": "2026-10-08T19:34:12.806Z",
"run_id": "family-gate-20261008",
"updated": "2026-10-08T19:35:58.408Z",
"evidence_record": {
"reason": "layer ablation is a research harness, not a release suite",
"at": "2026-10-08T19:34:12.806Z"
"requirement_id": "ROT-07",
"decision": "NOT RUN",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the acceptance-side ablation rows (layer 8 off against the same-crate windowed control: the draw cost, the floors' costs, the hot-set and 1.2x ceiling, the bit-R read; the mixer and shape axes' per-era invisibility and verifier rows), docs/analysis/class-v6/family-gate.md sections 6.4 to 6.12; the adversarial cost and the GPU setup rows per layer are the hash lane's and the adversary lane's",
"release_identity": {
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
},
"in_progress_since": "2026-10-08 18:3x UK",
"approvals": {
@ -3109,6 +3148,28 @@
"run_id": "team-2026-10-08",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": true
},
"harness:family-gate": {
"requirement_id": "ROT-07",
"decision": "NOT RUN",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the acceptance-side ablation rows (layer 8 off against the same-crate windowed control: the draw cost, the floors' costs, the hot-set and 1.2x ceiling, the bit-R read; the mixer and shape axes' per-era invisibility and verifier rows), docs/analysis/class-v6/family-gate.md sections 6.4 to 6.12; the adversarial cost and the GPU setup rows per layer are the hash lane's and the adversary lane's",
"release_identity": {
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
}
}
},
@ -3139,25 +3200,30 @@
"manual_page": 32,
"owner_lane": "node lane (a283f5f0d364ceef0)",
"run_status": "NOT RUN",
"evidence_path": "",
"run_id": "",
"updated": "2026-10-08T19:22:31.798Z",
"evidence_path": "docs/analysis/class-v6/family-gate.md",
"run_id": "family-gate-20261008",
"updated": "2026-10-08T19:35:58.408Z",
"evidence_record": {
"reason": "the no-new-rules counterfactual is a research harness",
"at": "2026-10-08T19:22:31.798Z",
"method": "static",
"requirement_id": "ROT-08",
"decision": "NOT RUN",
"reviewer": "",
"claim_impact": "",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the no-rescue control document (38,000 drawn eras of the parameter family through the acceptance, the bound arithmetic on the counts, the honest line of what the gate cannot see), family-gate.md sections 1 to 6; the competitiveness envelope and the economic half are ECO's",
"release_identity": {
"commit": "",
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
}
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
},
"approvals": {
"scope_approved": null,
@ -3185,6 +3251,28 @@
"run_id": "",
"evidence": "",
"in_progress": false
},
"harness:family-gate": {
"requirement_id": "ROT-08",
"decision": "NOT RUN",
"method": "native",
"cell": "harness:family-gate",
"manifest_sha": "ac86d7910",
"run_id": "family-gate-20261008",
"evidence": "docs/analysis/class-v6/family-gate.md",
"in_progress": false,
"coverage": "partial: the no-rescue control document (38,000 drawn eras of the parameter family through the acceptance, the bound arithmetic on the counts, the honest line of what the gate cannot see), family-gate.md sections 1 to 6; the competitiveness envelope and the economic half are ECO's",
"release_identity": {
"commit": "ac86d7910",
"lockfile": "",
"binary": "",
"network_object": "",
"activation": "",
"profile_hashes": ""
},
"claim_impact": "",
"reviewer": "",
"at": "2026-10-08T19:35:58.408Z"
}
}
}

View file

@ -0,0 +1,22 @@
{
"run_id": "family-gate-20261008",
"manifest_sha": "ac86d7910",
"method": "native",
"cut_tip": "class-v6 ac86d7910 for section 6.12 (the final tree, the packs' draw read back equal); 04442d9ca for 6.11; class-v5 8f481459 for 6.1 to 6.9; class-v6 542cd7028 for 6.10",
"evidence_dir": "docs/analysis/class-v6/logs",
"boxes": [
"build-1",
"build-2",
"build-3",
"build-4",
"build-6"
],
"cells": [
{
"cell": "harness:family-gate",
"status": "NOT RUN",
"evidence": "docs/analysis/class-v6/family-gate.md"
}
],
"note": "recorded by the steward for research lane D (a07a99a3788566af2); NOT RUN in progress: the coverage is partial on all three cases"
}

View file

@ -485,6 +485,24 @@
"coverage": {
"POW-07": "the branch ran and its claim failed: the FP32 branch costs the card more energy and widens the chip edge; the master's register row 12 reads FAIL, published; the branch is excluded and unreachable on master (the grep evidence stays)"
}
},
"harness:family-gate": {
"command": "IGNEUM_FAMILY_GATE=1 [IGNEUM_FG_CLASS=<class string>] igneum_pow_test --ignored family_gate_era_census (the class v5 crate plus docs/analysis/class-v6/logs/harness-*.diff; one TSV row per drawn era); attack-f8 census --seeds 64 --nonces 1048576 --control window --by-site against the same crate",
"box_class": "release (the family gate's top class on every box)",
"fixtures": [
"F0",
"F1"
],
"cases": [
"POW-02",
"ROT-07",
"ROT-08"
],
"coverage": {
"POW-02": "partial: the acceptance rule's measured rates on drawn classes (the per-candidate rejection, the attempt histogram, the exhaustion count against the independent-attempt arithmetic, the (c'') and (c''') floors' refuse rates and the pre-floor ratio spread per width), the per-site largest-bucket and index-bit statistics at the rule's own 2^20 sample with the seven known-failed seeds (p4, p8, p10, p34, p212, p225, p15), and the F8-form census at 2^20 on the F8 seed set where the attack-f8 mirror validates (0 of 96 warps mismatching); the semantics, termination and memory-bound half and the independent re-implementation are the suite's and the hash lane's; the effect-size limits are the family-gate.md thresholds, predeclared in its section 3",
"ROT-07": "partial: the acceptance-side ablation rows (layer 8 off against the same-crate windowed control: the draw cost, the floors' costs, the hot-set and 1.2x ceiling, the bit-R read; the mixer and shape axes' per-era invisibility and verifier rows), docs/analysis/class-v6/family-gate.md sections 6.4 to 6.12; the adversarial cost and the GPU setup rows per layer are the hash lane's and the adversary lane's",
"ROT-08": "partial: the no-rescue control document (38,000 drawn eras of the parameter family through the acceptance, the bound arithmetic on the counts, the honest line of what the gate cannot see), family-gate.md sections 1 to 6; the competitiveness envelope and the economic half are ECO's"
}
}
},
"not_run": {
@ -500,8 +518,6 @@
"ADV-07": "lifetime rows are the adversary lanes' models",
"ROT-03": "miner-voted bring-forward needs a vote harness on the fault network F4",
"ROT-04": "seed-selection resistance is the census harness (the class v6 invention lane), not yet in the matrix",
"ROT-07": "layer ablation is a research harness, not a release suite",
"ROT-08": "the no-new-rules counterfactual is a research harness",
"ECO-02": "F7",
"ECO-03": "F7",
"ECO-04": "F7",