igneum/docs/evidence.md

Igneum facts: every Igneum 2.0 claim, with its status

8 October 2026. The facts page restarts at Igneum 2.0: one row per claim, each traceable to a landed document in this repository, the path in the row. Every number carries its label (measured, modelled, synthesised, designed); nothing is restated from memory.

The six labels

Every row's Status cell opens with one of five served labels from the 2.0 plan: TEAM-REPORTED (measured or exercised by the project, procedure and version named), MODELLED (a model's figure, never a measurement), PROPOSED (a design or a decision, no code or chain behind it yet), PENDING (built, awaiting the chain event or the row that activates it) or EXCLUDED (a claim the project does not make). The status word follows it.

The public wording ladder, and what supports each rung: "designed for GPU competitiveness" (the objective only); "team-tested" (a procedure, a version and a result on the record); "independently reproduced" (an unaffiliated operator's run); "independently reviewed" (a scoped review with its unresolved findings); "top-tier contender" (comparative evidence plus the operator proposition); "category leader" (never claimed without sustained adoption, and never with a number). Today nothing on this site stands above team-tested, and every chip row is modelled.

Label	Meaning
designed	A decision in the design document, the plan or the specification. No code carries it, or the code is a stub
implemented	Code exists with test vectors, and the vectors pass. Not run as a measurement of the claim
activated	The code is live on a named chain by its activation height, and a node's own line says so. Not yet a measurement of the claim
tested by the team	The claim was measured or exercised by the project's own people and agents on a named machine, and the result is in a landed document
reproduced externally	Somebody outside the project ran the published command on their own hardware and got the published result
reviewed independently	Somebody outside the project, paid or not, read the code or the rule and published their finding

Five rules for reading the table:

1. Nothing on this chain has been reproduced externally or reviewed independently. Every row's last column says "none yet". Nobody outside the project has run a published command or published a reading of the code yet, so the first four labels are the ceiling today.
2. A status applies to the exact version or document in the row. An audit of one version never covers a newer one; when the version changes, the status falls back until the new version is tested, reproduced or reviewed again.
3. "Tested by the team" on one machine is one machine. The rows say which. Cards the project rents are the project's own measurement, as are the project's own rigs; neither is a reproduction.
4. The labels stay distinct and a claim never moves up a label without the artefact that label names (the table "What would move a row"). The public reference repository exists (the specifications, the igneum-pow crate, the simulators and the test material, at git.igneum.network); the full node, the miner and the proving code open later, so a row that cites them is tested by the team at most until they do.
5. A failure is a fact. An experiment that did not work is a row like any other, labelled by how it was measured, with the document that records it.

Versions in the table: "the release manifest" is site/release-manifest.json, served at /release.json and regenerated at the Igneum 2.0 devnet's first block; "node fork" commits are the node fork's, which is not in this repository's history until the node opens; a document path is this repository's.

The table

Every row carries its release evidence packet at the end of its Result cell: parameters, procedure, raw result, independent-reproduction status, review scope and unresolved limitations, with the source commit in the Version cell. The public mirror follows master; a link that answers 404 is a file not yet synced. An operations record with box names stays in the repository and is not public; its row says so.

#	Claim	Where it is made	Status	Version or commit	Reproducible test	Result, date, machine	Independent verification
1	The Igneum 2.0 devnet (igneum-devnet-4): genesis 7c36b833 stamped 8 October 2026, chain id 4465 (eth_chainId 0x1171), 18 decimals, every upgrade on from block zero (class v5, the era VDF, finality v3, calibrated fees, enforced proving); the release manifest (/release.json) carries its commits and is regenerated at its first block	This page; the live page; /build	TEAM-REPORTED; activated: the first block f3dc319c at DAA 0, mined at 17:14 UK on 8 October 2026 and accepted on the seed at 17:17	the release manifest (/release.json): node 4cdcc488 on release-2.0.0-node, igneum-pow 1c420786, digest be5f4068	the node's own start line and its first block (the node lane's read of build-1's seed)	8 October 2026, 17:14 UK: the first block f3dc319c at DAA 0 on node 4cdcc488; an earlier object of the same genesis minted at the wrong decimals and is dead Case: GOV-01	none yet. Packet: parameters the manifest's network block (genesis 7c36b833, digest be5f4068, 18 decimals); procedure the node's start line and eth_getBalance over the public RPC; raw result the first block f3dc319c and the miner's balance 2,535,047,024,800,000,000 wei after it; reproduction none yet; review scope none; unresolved an earlier object of the same genesis is dead and its chain is not served
2	Proof verification is enforced in consensus on the Igneum 2.0 devnet from block zero (Deliverable 5's prerequisite): the node's start line reads "consensus proof verification from DAA score 0 (verifier_in_consensus set)" and names the shard program id and the aggregator id; a block carrying a statement without a valid proof is refused	docs/plans/igneum-2.0.md D5; this page; the litepaper (Proving)	TEAM-REPORTED; activated: on from block zero, the chain running since 17:14 UK on 8 October 2026	node 4cdcc488 on release-2.0.0-node (the manifest); the program ids are the ELF manifest's (proving/igneum-prove/elf/manifest.json)	the node's own start line; the no-rescue exercise itself (D5) is Open	8 October 2026: on from block zero in the 2.0 devnet's object, the chain running since 17:14 UK; the earlier devnet ran with the rule off; no block yet refused for a false proof on the record Case: ZKP-01	none yet. Packet: parameters verifier_in_consensus set from DAA 0, the shard program id 0x2b1a81cb and the aggregator id 0x474678f3 (the ELF manifest); procedure the node's start line; raw result the line itself; reproduction none yet; review scope none; unresolved no hostile submission has yet been refused on the record (D5's exercise)
3	The 64-register window per lane costs a GPU under 1 percent of rate at stock, and at most 5 percent per load with the liveness chain	The litepaper (class v6); docs/plans/igneum-2.0.md D1 (the placed 64-register rows)	TEAM-REPORTED; tested by the team	docs/analysis/class-v6/connected-state.md section 4; docs/design/class-v6-rotating-family.md section 10.0e	The class v5 nvcc harness and the kit worker, both packs on the same card minutes apart, 250 batches of 2^24, nvidia-smi at 1 Hz, vectors PASS on every row (connected-state.md section 4); the per-load rows of the full chain against the base (class-v6-rotating-family.md 10.0e)	8 October 2026, rented RTX 5090 (575 W cap) and RTX 4090 (450 W cap) at stock: energy per hash +0.6 percent on the RTX 5090 and -0.9 percent on the RTX 4090, inside the run-to-run noise; under 1 percent of rate; 80 to 87 registers per thread, no spill (all measured). Per load: RTX 5090 16.7 nJ base, 17.6 nJ full chain; RTX 4090 26.0 nJ, 27.0 nJ (measured). The lock row on the project's own rigs is owed Case: GPU-03	none yet. Packet: parameters a rented RTX 5090 and RTX 4090 at stock, the liveness chain, class v4; procedure the hash lane's harness rows in connected-state.md section 4; raw result energy per hash +0.6 percent and -0.9 percent, at most 5 percent per load; reproduction none yet; review scope none; unresolved two cards, one day, no AMD or Apple row
4	Reorganising the same work around live state (the connected-state variant, experiment D2(a)) does not reduce a specialised chip's edge: KILL as a class	docs/plans/igneum-2.0.md D2(a); this page	TEAM-REPORTED; tested by the team (a published failure)	docs/analysis/class-v6/connected-state.md (the verdict, section 6)	The census, liveness and GPU rows in connected-state.md sections 2 to 4; the chip side priced on the drawn program by synthesis (a model, never a lower bound)	8 October 2026, verdict 17:25 UK: the window is necessary (63 of 64 registers live at every address, measured) but only its width reaches the chip, +1.2 pJ per lane-op at N5 (synthesised); the window moves the chip's edge 1.10x node for node against a 1.25x gate (modelled); the GPU side +0.6 percent energy per hash on the RTX 5090, -0.9 percent on the RTX 4090 at stock (measured). Rearranging the dependency graph of the same operations moves neither side Case: POW-04	none yet. Packet: parameters the connected-state generator variant behind a flag, the census TSVs; procedure connected-state.md; raw result +1.2 pJ per lane-op at N5 on the chip side, the GPU side unmoved; reproduction none yet; review scope none; unresolved the chip side is modelled, not measured
5	"A GPU-secured network for Ethereum-compatible applications and verifiable computation." served on every page	Every page	PROPOSED; designed (served)	docs/plans/igneum-2.0.md (the objective: the positioning line)	node tools/ci/ledger-text-check.mjs: the sentence pinned (R0) on the home page, the litepaper and this page	8 October 2026: on this page; the home page and the litepaper carry it as their 2.0 text lands (designed) Case: GOV-08	none yet. Packet: parameters none (a served sentence); procedure the ledger-text check pins it on every page; raw result the check's count; reproduction not applicable; review scope the three external reviews the founder accepted; unresolved none
6	The chip claim as served: "Igneum remains competitive on accessible commodity GPUs even when specialised mining hardware is assumed to exist, remain compatible and seek profit; its security does not rely on identifying that hardware or retiring it through emergency changes." Under it the three statements, separate: energy ("Current modelling estimates a 1.5x to 3.1x energy-efficiency advantage for the specialised designs assessed as complete machines against the GPU tier, from a board on commodity DRAM at 1.5x to an SRAM-store die at 3.1x (1.5x to 2.3x on the GPU's own node). For the board on commodity DRAM the two independent chip models agree after placement: 1.5x to 2.1x as a complete machine (the 8-lane 18-family core at the low end, the placed 32-lane floor at the high end) and 2.0x to 2.4x for the board alone on the GPU's own node; a node ahead 1.8x to 2.4x and 2.45x to 2.8x; the stored-half hybrid 1.9x to 2.6x on the same node and 2.4x to 3.2x a node ahead (both placed and routed on ASAP7, the node scaling claimed, the machine terms modelled). The standard's P04 target as approved: R_E at most 1.5 on the same node and one node ahead; today's placed bracket reads against it as a FAIL to work against." Per machine on the same node and a node ahead: the DRAM board 1.5x to 1.6x and 1.8x, the hybrid 1.9x to 2.1x and 2.4x to 2.6x, the die 2.3x and 3.1x; across the adversary's lane-count choice the same-node bracket is 1.5x to 2.1x and a node ahead 1.8x to 2.6x; 3x to 6x per dollar of hardware at list price. MODELLED: the GPU side measured, the chip's core placed and routed, the rest of the machine modelled, no chip measured, hardware cost approximate within 2x), economic and response capability, rotation an optional improvement. Class v5 derives the dataset from chain state; whether that excludes a specialised design is under evaluation (Deliverable 3), since a design that tracks state is not excluded by staleness. The dataset policy (decided 8 October 2026, 20:05 UK: the genesis size is 4 GiB, the lane's recommendation taken; 2 GiB and 5.5 GiB the costed alternatives beside it; later steps at 5 percent): "The dataset’s size is a one-off hardware ticket on specialised designs and a running energy tax on commodity cards: at 4 to 5.5 GiB a locked Blackwell card pays 12 to 14 percent more energy per hash than at 1 GiB (an 8 GB AMD card under 3 percent of rate), while the board on commodity DRAM pays nothing and the SRAM designs pay a hardware cost that does not change their outcome. So the genesis size is 4 GiB on that scoring (decided), and every later step is scored at 5 percent against it and taken only when the chain’s own state outgrows the dataset, not on a calendar." (decided; the energy-against-size curve on the litepaper: 1, 2, 4 and 5.5 GiB at stock and at the lock measured on an RTX 5090 against its own 1 GiB control, +0.7 and +4.3 percent at stock, +11.7 and +13.4 at the lock for 4 and 5.5 GiB, the 2 GiB knee cell a paired read in flight; the exclusions by memory; the SRAM tickets modelled). A failure, reported as found: "ECO-05, the standard's coexistence sweep, was run on a register frozen before any result against a sourced revenue reference of USD 31.65 million a year of miner revenue (Ethereum Classic's trailing year at its current era 6 reward) and FAILS the envelope as the chain stands. Of sixteen revenue and tariff worlds (a quarter, one, four and ten times the reference; electricity at 3, 10, 25 and 40 cents per kilowatt-hour) only one sustains new entry across two vendors, ten times the reference at 3 cents (about USD 316 million a year), because on this hash the measured AMD and Intel cards cost four to six times a Blackwell card per joule and never earn a new entrant's purchase back at list price below that world; in that one world no specialised design sits inside the envelope against the whole cohort (the board on commodity DRAM is within the envelope only against the best Blackwell card, the die and the hybrid against no card of today's), and the quarter-reference worlds at 25 and 40 cents are collapse worlds with no rational profitable operator. What moves the result is the specialist's hardware cost per unit of work, the honest cards' own efficiency on this hash, above all the AMD and Intel cards' energy, and the dataset's size floor; not a smaller network, a higher token price or any chip's death. Reported as found; the register, the full result cube and the model are published for independent reproduction." (Labels: the register frozen at 18:37 UK on 8 October 2026, the reference corrected by a second public read and the run repeated with the verdict unchanged, the cube MODELLED on measured card rows, two cohort rows BLOCKED, the RX 7600's knee and the Arc's watts, their measurement running tonight; the registry row ECO-05 on /acceptance; the results file docs/analysis/class-v6/eco-05-results.md.) The energy ratio is not the pass criterion: the coexistence model (docs/analysis/class-v6/coexistence-model.md) is, and its first run's result is served with its conditions: A specialised supplier may earn a normal return; ordinary GPUs remain sufficiently close in total cost, widely obtainable and useful outside mining that new operators can still compete. On today's modelled rows that holds for the chip anyone can build, a stored-dataset board on commodity DRAM, at a productive life of one to three years: its cost per accepted unit of work sits within the range of the best GPU owner and entrant, it passes six of the seven conditions (a third of the network in boards now costs less than a year's revenue at the final hardware price), and a fleet of it holds a minority of the network with GPU entrants still setting the price. For the SRAM-store die the outcome turns on its hardware cost per unit of work, not its energy advantage: at the reconciled machine cost, set by the power train and the shadow core rather than the die, it fails the cost, hardware, fleet and margin conditions at every point of the band, a modest fleet holds about two fifths of a growing network and three fifths of a flat one on arrival and takes every flat or shrinking network within five years, and the only coexistence-shaped outcome is private supply in a growing network; what holds it is the investment decision, since its economics are project economics. A third design, a DRAM board with the hottest half of the dataset in on-board SRAM, sits between the two: it coexists only in a growing network at cheap GPU electricity and fails the cost conditions in a flat or shrinking one, and the dataset's size floor is a real lever on it where it was none on the SRAM die. What holds the die is the investment decision, not the hash; every chip figure here is modelled, not measured, and the chip's hardware cost per unit of work is approximate within 2x.	The litepaper (the chip model)	MODELLED; designed (the bracket modelled; the GPU side tested by the team)	docs/design/class-v6-rotating-family.md section 10 (10.0h to 10.0n, 8 October 2026); docs/plans/igneum-2.0.md D3 and D4	the scoring rule in the close (the minimum over workloads of the maximum over free adversarial designs of the GPU's joules per hash over the adversary's, under the 10 percent GPU-cost budget at the lock, the verifier limit, cross-vendor correctness and hardware accessibility); the placed rows are D3's	the GPU side measured: the RTX 5080 at its 1,100 MHz lock 2.06 microjoules per hash and the RTX 5090 at its 1,300 MHz lock 2.33 (8 October 2026, the project's own rigs and rented pods); the chip side synthesised and claimed, its placed gated row pending Case: ADV-01	none yet. Packet: parameters the placed full 18-family core on ASAP7, claimed node scaling, the GDDR7 board, the RTX 5090 at its 1,300 MHz lock on class v4; procedure class-v6-rotating-family.md section 10 and coexistence-model.md; raw result the energies and verdicts as served; reproduction none yet; review scope three external reviews of the close, findings taken; unresolved no chip measured, hardware cost approximate within 2x, the k lane's 32-lane placed row pending
7	Mining and proving together on one card needs a 16 GB card at the proposed 5.5 GiB dataset (the genesis size is 4 GiB, decided 8 October 2026; at 4 GiB the miner holds less and the rule is not yet measured there): the miner holds about 6.1 GiB and a compressed shard proof peaks at about 7.5 GiB, so 8 GB and 12 GB cards time-share (the app pauses the miner for the proof). NVIDIA proves; AMD and Apple mine	The litepaper (Proving, vs RandomX); the miner page	TEAM-REPORTED; tested by the team	docs/analysis/class-v6/coexist-rows.md (the 5.5 GiB ds55 miner beside igneum-prove-host-0317, compressed at threshold 2^26, the served sm_86 and sm_89 floors)	the RESULT rows in that file, verbatim from the runs	rented RTX 3060 12 GB: 26.82 MH/s at 117.4 W with 6,129 MiB resident; the compressed shard 13.2 s, peak 7,525 MiB (6,129 + 7,525 = 13,654 MiB against 12,288); rented RTX 4060 8 GB: 18.84 MH/s, 6,116 MiB resident; the shard 8.2 s, peak 7,532 MiB; 8 October 2026 Case: GPU-05	none yet. Packet: parameters the 5.5 GiB ds55 miner, igneum-prove-host-0317 compressed at threshold 2^26, sm_86 and sm_89; procedure coexist-rows.md; raw result the RESULT rows verbatim; reproduction none yet; review scope none; unresolved two cards on one day, the 6 October rows stand as 1 GiB-dataset rows
8	Finality is active on the Igneum 2.0 devnet from checkpoint 241, 115 minutes after block one, with 78 percent of active weight signing	The live page's strip; the explorer; this page	TEAM-REPORTED; activated: the first lock at 19:08:25 UK on 8 October 2026 (checkpoint 241, DAA 7,247, 77.8 percent of active weight; checkpoint 242 at 19:08:48)	the node's checkpoint state through the observer (/api/live finality.active, latest_locked_index); node 4cdcc488 on release-2.0.0-node	the observer's finality fields read at the edge (supported true, active true, latest_locked_index 244 at 19:10 UK). Case: FIN-01	8 October 2026, 19:08 UK: the first lock 115 minutes after the first block at 17:14; the weight window filled at DAA 7,200. Packet: parameters rule v3 from checkpoint DAA 0, the 7,200 DAA presence window, two thirds of active and of total weight; procedure the observer's checkpoint table; raw result checkpoint 241 locked at DAA 7,247 with 77.8 percent of active weight; reproduction none yet; review scope none; unresolved a pause has not yet been observed on this chain	none yet

Count by status

Status	Rows
designed	2 (rows 5, 6)
implemented	0
activated	3 (rows 1, 2, 8)
tested by the team	3 (rows 3, 4, 7)
reproduced externally	0
reviewed independently	0

8 rows. The rendered page is site/evidence.html (served at /evidence), generated from this file by site/build.mjs; the text is judgement, so this file is edited by hand and the page follows.

What would move a row

From	To	What it takes
designed	implemented	Code with test vectors that pass
implemented	activated	A node's own start line on a named chain naming the rule and its activation height
implemented	tested by the team	A landed document with the machine, the date, the command and the number
tested by the team	reproduced externally	The code the row names public in the reference repository (the specifications, the pow crate, the simulators and the test material are; the full node, the miner and the proving code open later), the command published, and a third party's run with the same result, linked from the row
reproduced externally	reviewed independently	A named reviewer's published finding on that version. Funding for review is docs/plans/funding.md
any	the row's status falls back	A new version of the code or rule the row names