igneum/docs/evidence.md

Igneum evidence: every public claim, with its status

3 October 2026. One row per claim the homepage (site/index.html) and the litepaper (site/litepaper.html) make. Every number here is copied from docs/bench-log.md, which names the machine, the date and the command; nothing is restated from memory. Where a bench-log figure is approximate, this table says so.

The five labels

Label	Meaning
designed	A decision in the design document or the specification. No code carries it, or the code is a stub
implemented	Code exists in this repository with test vectors, and the vectors pass. Not run as 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 the bench log
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

Three 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". The repository is private until January 2027 (site/journey.json), so the first two labels are the ceiling today.
2. A status applies to the exact version in the row. An audit of one version never covers a newer one; when the version changes, the status falls back to "tested by the team" until the new version is reproduced or reviewed again.
3. "Tested by the team" on one machine is one machine. The rows say which. Discrete AMD, Intel and a 2019-class CPU core have not run anything.

Versions in the table: igneum-pow is the Rust crate at igneum-pow/Cargo.toml version 0.1.0. "Repo" commits are this repository's. "Fork" commits are vendor/igneum-node and its worktrees (-r3, -diff, -exec, -harness), which are not in this repository's history; the row names the fork commit by its message as the bench log does. The spec is docs/spec/ version 0.1.

The table

#	Claim	Where it is made	Status	Version or commit	Reproducible test	Result, date, machine	Independent verification
1	A new mining program every hour, compiled by the miner, with no human in the loop	Homepage hero and "This hour's mining program"; litepaper Mining	tested by the team	igneum-pow 0.1.0; repo 9812466; fork "PoW: header-bound lottery engine, real-hash miner modes, genesis bits 0x1e400000"	igneum-miner mine --engine igneum-pow against a 3-node igneumd --devnet, with proto-metal/igneum-bench --serve as the GPU worker; bench-log "first devnet blocks on the real lottery hash"	Epoch change crossed live at DAA 3,600: new seed, a 128-load program compiled by the Metal worker in 129 ms, 0 rejected blocks across the change. 3 October 2026, Apple M5 Max. The epoch seed on the devnet is the epoch block hash; the VDF of row 2 is not wired in yet	none yet
2	The program seed passes through a 10-minute verifiable delay from a certified checkpoint, so nobody can grind the seed	Litepaper Mining, vs RandomX ("Closed by a verifiable delay")	implemented	repo 792776e; proto-vdf/	proto-vdf full 10-minute runs and the tamper cases in proto-vdf/README.md; bench-log "proto-vdf"	Class group 1024-bit: 163,000 squarings per second, 10-min eval 585.4 s, prove 9.1 s on 12 threads, verify 4.47 ms, 516-byte proof; wrong checkpoint, flipped seed bit and T+1 all rejected; grinding model gains 0 blocks per epoch with the delay against +3.62 at a 30% advantage without it. 3 October 2026, Apple M5 Max, one core. Prototype only: not in the node, not reviewed against chiavdf (O-4.1)	none yet
3	The dataset is memory-hard: computing an item costs more than loading it	Litepaper Mining and vs RandomX; homepage vs RandomX ("Memory 2 GB, growing")	tested by the team	repo 58a5a63; igneum-pow 0.1.0 (memhard.rs); proto-metal/MEMHARD.md	proto-metal/igneum-bench --inline-dataset against the honest run at 1 GiB and 256 MiB; bench-log "memory-hard dataset"	Honest 45.2 Mhash/s, inline (never reads the dataset) 9.49 Mhash/s, ratio 0.21, at 1 GiB; 0.10 at 256 MiB. 3 October 2026, Apple M5 Max. Apple only: the shortcut ratio has not run on NVIDIA or AMD (O-1.5). Review round 3 priced a 256 MiB on-die cache chip at about 2.4x, approximate, which the measurement does not answer (ledger M16)	none yet
4	The same program produces identical hashes on three GPU vendors, cache and dataset included	Litepaper vs RandomX ("Bit-exact on Apple and NVIDIA, measured"), For miners; homepage	tested by the team	repo f2e903e, 0f1fdaf; pack proto-cuda/packs/igneum-genesis-mh; igneum-pow 0.1.0	The 96 test vectors of the pack through proto-metal/igneum-bench, proto-cuda/host.cu, proto-opencl/host.c; batch fingerprint at --batch-log2 24; bench-log entries "RTX 5090, memory-hard dataset", "AMD gfx1036", "RTX 5090 through NVIDIA OpenCL"	96/96 on Apple Metal (M5 Max), NVIDIA CUDA and NVIDIA OpenCL (RTX 5090, Windows), AMD OpenCL (Ryzen 7 9800X3D integrated gfx1036, 1 compute unit), Apple OpenCL, pocl and two CPU references; batch fingerprint 98af644e993239e2 over 16.7 million nonces identical on the AMD chip and the 5090. 3 October 2026. The AMD device is an integrated chip; no discrete AMD card and no Intel card has run anything (O-1.15)	none yet
5	A CPU verifies one hash in under 10 ms by simulating one warp	Litepaper Mining ("about ten milliseconds"), vs RandomX; roadmap gate 2	tested by the team	repo 75cac18; igneum-pow 0.1.0 (verify.rs)	cargo test and the crate bench in igneum-pow/; bench-log "igneum-pow: Rust crate bit-exact with proto-metal"	0.411 to 0.579 ms per 32-lane warp steady, 0.41 to 0.87 ms cold, average of 20, 1 GiB dataset, cache held, one M5 Max performance core; the Swift verifier 0.63 to 1.2 ms. Gate margin about 17x on this core. 3 October 2026. Not measured on a 2019-class laptop core (O-1.14)	none yet
6	The hash is bound to the header: one nonce serves one header, and a wrong nonce is rejected	Spec 1.6; litepaper Mining (implied by "checks a hash")	tested by the team	repo 33f7b33, 9812466; igneum-pow 0.1.0 (bind.rs, 8 bound vectors, 29 crate tests)	igneum-miner bad-nonce against a devnet node; igneum-pow hash-bound for the 96-nonce job across the 2^32 lane boundary; bench-log "first devnet blocks on the real lottery hash"	833 blocks accepted by igneum-lottery-v1-bound on 3 nodes, 0 rejections; bad-nonce gave Reject(BlockInvalid); Metal, OpenCL and CUDA (emulated) workers bit-exact with the crate on the lane-boundary job. 3 October 2026, Apple M5 Max	none yet
7	The devnet runs at one block a second	Homepage stats ("1 / s"); litepaper Speed; roadmap phase 3	tested by the team	repo 9812466, e9328c6; fork worktree vendor/igneum-node-diff branch difficulty	3-node CPU devnet, igneum-miner mine --engine igneum-pow, 660 s; the dual-lane rule's 3-node test network (ports 26800 to 26821); bench-log "first devnet blocks" and "difficulty controller"	CPU devnet: 1.29 blocks/s over 641 s, 1.03 blocks/s over blocks 610 to 816 after the first retarget, sink identical on 3 nodes at 61 of 64 samples. Kaspa's sampled rule on the overnight devnet did not hold the rate across a hashrate step (5.44 blocks/s for five minutes, 4.7x the schedule for eight minutes). The Igneum dual-lane rule's test network reached 1 block/s within 10% after 132 s, worst gap 7.3 s. 3 October 2026, Apple M5 Max. The phase 3 gate also asks for proofs under 60 s behind the tip; no proof exists (row 15)	none yet
8	Blocks are mined by GPUs on Apple and NVIDIA	Homepage live strip; journey phase 3 ("GPU miners on three vendors")	tested by the team	repo 9812466, e9328c6; fork as row 1	Metal worker proto-metal/igneum-bench --serve driven by igneum-miner --worker, 300 s; the overnight devnet record sim/difficulty/devnet-2026-10-03.csv; bench-log "first devnet blocks" and "difficulty controller"	Metal: 506 jobs, 5,636 blocks found and accepted, 0 rejected, 0 CPU/GPU mismatches, 28.2 MH/s wall. NVIDIA: the overnight devnet record shows the PC's RTX 5090 mining at 116 MH/s estimated from the blocks, and the finality follower counted eight RTX 5090 identities at 862 to 936 blocks each. 3 October 2026, Apple M5 Max and the Windows PC	none yet
9	Blocks are mined by a GPU on AMD	Journey phase 3 ("three vendors")	implemented	repo 0f1fdaf; bound kernel kernel_bound.cl in the pack	proto-opencl/host.c --serve on an AMD device against a devnet node	The bound OpenCL kernel is bit-exact with the crate on Apple OpenCL and the memory-hard pack passes 96/96 on AMD gfx1036, but no block count mined by an AMD device is recorded in the bench log. Until one is, the three-vendor mining claim is two vendors mined plus one vendor verified	none yet
10	Checkpoints lock every 30 s of chain at two thirds of all 30-day weight (raised from 56.7% of total on 4 October 2026), and the floor stops conflicting locks in partitions and eclipses	Litepaper Finality, "What Igneum does not claim"; homepage "locked every 30 seconds"	tested by the team	repo bbb264a (simulation), 60b1412 (fork run); fork "Finality: BLS12-381 vote keys ..." through "Miner: BLS identity ..."; spec 3.10	sim/finality_v2.py (results in sim/results_v2.md); the 4-miner test network igneum-devnet-7 with getFinalityCheckpoints on three nodes; bench-log "finality rule V2" and "igneum-node devnet v2"	Simulation: 0 conflicting locks in every honest partition and eclipse scenario with the floor; without it both sides of a 50/50 split lock after 60 min. Test network: 72 of 72 determined checkpoints locked on all three nodes, lock latency median 0.80 s, p90 1.08 s, 0 conflicting certificates; an equivocating key stripped at index 43 and excluded; with one voter left (39.6% of total) 0 locks for 749 s, then the heal locked 13 checkpoints within 30 s. 3 October 2026, Apple M5 Max, 53 minutes. Not on the live devnet (its miners do not vote yet); the simulation has no DAG; one unexplained stall of all three nodes in the first run, not reproduced	none yet
11	Hashrate that arrived today has almost no vote: ten days of the whole network's hashrate to reach a third of the weight, twenty for two thirds; 51% never reaches two thirds while honest miners stay	Litepaper Finality; homepage firsts	tested by the team	repo bbb264a; sim/finality_v2.py	Scenario B of sim/finality_v2.py, seeds 7 and 11	share(t) = (t/30) x a/(1+a) holds to 0.04 points; a renter equal to the whole honest network (a = 1) crosses 1/3 on day 20 and never reaches 2/3; a = 9 crosses 1/3 on day 11.1 and 2/3 on day 22.2. The ten-day figure is a = infinity, honest miners gone. 3 October 2026, Apple M5 Max. A model with 1,000 Pareto keys and no DAG; the harness scenarios against the real node (1b, 3b, 4b) are stubs	none yet
12	The difficulty rule recovers from a hashrate step within about a minute, where Kaspa's sampled rule never settles	Spec 2.3; litepaper Speed (implied); bench page	tested by the team	repo e9328c6; fork worktree vendor/igneum-node-diff branch difficulty; sim/difficulty/sim.py	sim/difficulty/sim.py on nine profiles plus the devnet record; the 3-node test network with "difficulty_rule" in the override file; cargo test -p kaspa-consensus --lib difficulty (9 pass); bench-log "difficulty controller"	Simulator, settled seconds: x50 step 62 (Kaspa 1,542), /50 step 657 (Kaspa 12,296), the devnet's 75x step 79 (Kaspa never). Test network: within 10% of 1 block/s after 132 s warm-up, 214 s on a join, 85 s on a leave. 3 October 2026, Apple M5 Max under load 50 to 98. Monero and LWMA baselines reproduced from memory, approximate; no DAG in the simulator; harness scenarios 2b and 7b are stubs	none yet
13	Every node executes the ordered transactions natively and reaches the same state root	Litepaper Proving ("Every node executes ... natively"), Building ("runs on Igneum unchanged")	tested by the team	repo f5f8c80; fork worktree vendor/igneum-node-exec branch execution-layer; revm 43.0.3	node tools/evm-smoke/smoke.mjs against a 3-node igneumd --simnet; igneum-exec-diff seq.json; bench-log "execution layer devnet v3"	87 of 87 viem checks; state roots identical on 3 nodes at chain blocks 0, 56, 74 and 78; 57 executed transactions and 19 skipped copies agree with plain revm, 0 mismatches; balances match receipts to the wei. 3 October 2026, Apple M5 Max. Simnet skips proof of work, the prover is a stub, state is rebuilt from genesis at start, no EVM transaction relay between nodes	none yet
14	Ethereum bytecode runs unchanged, with the documented differences of spec 7.1	Homepage Build card; litepaper Building	tested by the team	as row 13	tools/evm-smoke/smoke.mjs: deploy via viem, increment, hashLoop, eth_estimateGas, eth_getLogs	Deployment, calls, reverts, logs and gas estimates behave as viem expects; chain id 4463; the prototype pgas table gives 0.0095 to 0.028 pgas per gas, below the design's band before calibration. 3 October 2026, Apple M5 Max. The Prover precompile, proof records and the shard planner are not implemented	none yet
15	Every block is proven, with the proof landing within about a minute at launch	Homepage stats ("~60 s to a proof"); litepaper Proving; roadmap phase 3 gate	designed	spec 7.2; docs/design/execution-layer.md section 5	None exists. The phase 2 benchmark standard is docs/benchmarks/proving-e2e.md	No SP1 shard has been proven on any card in this repository (ledger P1, P3). The devnet prover is a stub that signs claims. The 60-second figure is a design target	none yet
16	A 12 GB card proves one shard in about 20 s	Litepaper Proving ("The proving budget"); roadmap gate 2	designed	spec 5.1 (Target)	Replaced as a gate by the end-to-end standard in docs/benchmarks/proving-e2e.md: fixed workloads, job-to-accepted-proof latency, no growing backlog	Unmeasured. A per-shard time can be met by shrinking the shard, so the project no longer uses it as a pass mark	none yet
17	The chip resistance target: a chip gains under 2x over a GPU	Litepaper Mining, "What Igneum does not claim"; homepage "no chip can be built for it"	designed	spec 0.2 (Target); O-1.17	Public benchmark with a leaderboard by card model and a standing bounty, January 2027 (O-1.17); the on-die-SRAM test on the RTX 5090 (R3.5)	A target, not a measurement. Review round 3 priced a recompute chip with the 256 MiB cache on die at about 2.4x, approximate, before the usual chip-versus-GPU integer gain; the design answer (cache larger than any die) is open (spec 1.16)	none yet
18	The chip resistance measurements: the program is random-access bound, not bandwidth bound, and sits beyond a card's on-chip cache	Litepaper Mining ("bound by memory bandwidth", to be corrected), vs RandomX "Measured so far"	tested by the team	repo aba248d, f2a1a64, 4b95c5e	RTX 5090 dataset sweep 4 MiB to 1 GiB with proto-cuda/host.cu; bench-log "RTX 5090 first run" and "dataset sweep"	At 1 GiB: 228.1 Mhash/s, 23.7 G random loads/s, 94.9 GB/s useful against a 1,638 GB/s dataset fill; inside the 96 MiB L2 (4 and 64 MiB) 1,340 to 1,353 Mhash/s, about 5.8x faster; 104 against 128 loads per hash gives 228 against 185 Mhash/s, proportional. 3 October 2026, RTX 5090, Windows, CUDA 12.8. Prototype dataset 1 GiB against 2 GB at genesis; a pure random-read microbenchmark (R3 chip designer, attack 2) has not run	none yet
19	The lottery hash is sound as a hash: uniform output, deterministic, no out-of-bounds read, fuzzed	Litepaper vs RandomX ("Every number above is measured and logged")	tested by the team	repo c52307e, 58a5a63; proto-metal/TESTS.md	proto-metal/igneum-bench --fuzz --edge --stats --determinism --memcheck; bench-log "hardening tests" and the re-run on the memory-hard dataset	10,200 random programs, 1,305,600 hashes, 0 mismatches; 14 of 14 edge cases; bit frequency within 2.90 sigma, avalanche mean 31.99 to 32.04 of 32; deterministic fingerprint across 5 runs; every dataset read masked. 3 October 2026, Apple M5 Max. Statistics are not a security proof; the weak-program census (O-1.3) and the seed derivation review (O-1.4) are open; the fuzz set has run on Metal and the CPU only	none yet
20	No premine, no pre-sale, no allocation: every coin is minted by the schedule and every coin goes to the block producer (80%) and the proving pool (20%)	Homepage stats and Economics tiles; litepaper Supply, Economics	implemented	repo 6ac80a3; fork "igneum-node devnet v0"; consensus/core/src/igneum.rs, coinbase.rs	cargo test -p kaspa-consensus-core igneum (8 pass: subsidy table, ramp, split, cap) and cargo test -p kaspa-consensus coinbase (8 pass); igneum-miner inspect 40; bench-log "igneum-node devnet v0"	Coinbases on the devnet: 80/20 exact on 39 of 39 single-payee blocks, the 20% to the igneum-proving-pool-v0 output; the per-second schedule sums to under the 4,000,000,000 cap by less than 100 coins; 3,168,808,781 units per DAA second in years 0 to 2, halving at 63,115,200 DAA s. 3 October 2026, Apple M5 Max. The devnet genesis carries no allocation; the mainnet genesis does not exist yet, so the claim is about the code and the stated rule, not a launch that has happened	none yet
21	The proving pool's 20% reaches shard provers and aggregators	Litepaper Economics; homepage "20% provers"	designed	spec 5.3	None. The pool output exists (row 20); the payout from it against proof records is unwritten	The escrow accumulated on the simnet (92.55 IGN at the end of the v3 run) and nothing can draw it. Rule decided: per block, divided among shards by consensus proving cost, sortition to 8 provers for 10 s then open (spec 7.2)	none yet
22	The base fee is burned in full and the priority fee splits 80% to the miner and provers, 20% to the apps whose code ran	Homepage Economics caption and Build card; litepaper "Where fees go"	tested by the team	repo f5f8c80; fork worktree vendor/igneum-node-exec	tools/evm-smoke/smoke.mjs receipt checks; bench-log "execution layer devnet v3"	Transfer receipt: burnedProvingFee 200 gwei, minerTip 16,800 gwei (80%), unregistered developer share 4,200 gwei burned; contract call: 80% to the miner, 20% credited to the payee the constructor registered, balance delta equal. 3 October 2026, Apple M5 Max simnet. The provers' part of the 80% is not split out (no provers exist); the base fee stayed at the 1 gwei floor throughout	none yet
23	No fee to any team, foundation or fund; 0 admin keys in consensus	Homepage Economics tiles and caption; litepaper "No fund, no foundation" and Governance	designed	spec 5.5, 5.6 (decided 3 October 2026); spec 08	Reading: no coinbase output, fee route or consensus key in the fork names any party (coinbase.rs, docs/fork-divergence.md)	The emission code has two outputs (row 20) and the fee code has three routes (row 22), none to a team. The 1% fee of the official client is a client setting, not a protocol rule, and is not implemented (no client exists). The release key of spec 08 signs client updates and holds no consensus power; its custody policy is open (O-8.1)	none yet
24	External proving jobs pay 90% to the provers who delivered and burn 10%, once settled in IGN	Homepage "IGN burned from jobs, phase two"; litepaper Proving and Economics	designed	spec 5.4	None. Needs the proof bridge (spec 7.3, phase two) and the settlement switch (O-5.2)	At launch jobs are paid on the customer's chain in the customer's currency and nothing is burned (ledger P10). No job market code exists	none yet
25	The 4 billion cap, halving every two years, with a 30-day ramp from 10%	Homepage "4B IGN hard cap"; litepaper Supply and the emission chart	tested by the team	repo 6ac80a3; fork consensus/core/src/igneum.rs	cargo test -p kaspa-consensus-core igneum; bench-log "igneum-node devnet v0"	Ramp day 0 paid 10.03% of the full rate (317,767,704 units at DAA 806); the schedule table and the cap assert in the crate's own tests. 3 October 2026, Apple M5 Max. Base unit (8 or 18 decimals) is open (O-2.6); the spec was changed to follow the code's 365.25-day year (ledger E9) and a test that reads the published numbers back is still owed	none yet
26	A phone or browser verifies the chain from a locked checkpoint, at about 3.44 MB per day in checkpoint mode	Homepage "Browser checks Igneum" card (preview, commit f874f80); litepaper Building ("Light clients"), firsts row 6	designed	spec 10 (10.5 bytes per day: 3.44 MB at 1,000 voters, 6.68 MB at 10,000, derived, approximate); repo f874f80 for the browser card	None for the byte figure; site/verify/ for the card. BLS verification on a phone and in WebAssembly is O-10.3; the full-header mode on a phone is O-10.4	The homepage card verifies the latest certified checkpoint's BLS certificate in the tab against a voter list from the node (light client v0, 3 October 2026). The byte figure is arithmetic on designed sizes (header 400 bytes, proof 400 bytes), measured nowhere; the proof the card would check does not exist (row 15)	none yet
27	The node survives malformed input, floods, withholding, partitions and eclipses	Litepaper Speed ("GHOSTDAG, the BlockDAG consensus proven on Kaspa"); spec 2	tested by the team	repo 394030c; fork worktree vendor/igneum-node-harness; tools/harness/	tools/harness/ scenario runner against a private igneumd test network (ports 27200+); bench-log "consensus attack harness"	63 malformed cases, node up on every one; timestamp bounds exact; withholding at 10%, 25%, 33% and 45% released every 5 blocks within 2 sigma of share; partitions of 120 s to 3,700 s healed to one chain in 10 s; eclipse victims rejoined in 10 s; 50x template, submit and mempool floods left template p95 under 4 ms. One FAIL: a 45% withholder releasing every 20 blocks took 50.7% of blues (bound 47.4%). 3 October 2026, Apple M5 Max, load 50 to 61. Finality and difficulty scenarios are stubs until those branches merge	none yet
28	Headers are validated cheaply before the lottery engine runs, so forged timestamps cannot force 256 MiB cache builds	Spec 2.4; ledger M15	tested by the team	repo 0953ec7; fork worktree vendor/igneum-node-r3 branch r3-fixes at 5166ee26	measure_m15_attack_before_and_after (ignored test, release, --features igneum-pow); kaspa-pow 8, header_processor 1, p2p pow_guard 2 tests	50 forged headers: before, 50 cold builds in 10,595 ms and the live day evicted; after, 0 builds, all 50 rejected in 14 ms. 3 October 2026, Apple M5 Max under load 60 to 110. Measured through the validate path with skip_proof_of_work, not the daemon RPC; not merged into the main fork branch	none yet

Count by status

Status	Rows
designed	7 (rows 15, 16, 17, 21, 23, 24, 26)
implemented	3 (rows 2, 9, 20)
tested by the team	18 (rows 1, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 18, 19, 22, 25, 27, 28)
reproduced externally	0
reviewed independently	0

28 rows. The rendered page is site/evidence.html, kept in step by hand with this file; the bench page is generated, this one is not, because its text is judgement, not a log.

What would move a row

From	To	What it takes
designed	implemented	Code in this repository with test vectors that pass
implemented	tested by the team	A bench-log entry with the machine, the date, the command and the number
tested by the team	reproduced externally	The repository public (January 2027), 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