igneum-network/igneum
1
0
Fork 0
Code Issues Pull requests Releases Activity

Evidence row 30: four machines on the app

Browse source 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs 2026-10-04 13:56:11 +00:00
parent 1446d012cf
commit c8b026b8d9
4 changed files with 24 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

25
app/igneum-app/src/engine.rs
View file

@ -349,6 +349,9 @@ pub struct Engine {
last_upload: Instant,
upload_thread: Option<std::thread::JoinHandle<()>>,
ota: crate::ota::Updater,
jobs: crate::jobrun::Jobs,
/// a remote job has the miners stopped; they restart when the runner lets go
job_hold: bool,
stamp: String,
label_base: String,
wrapper: bool,
@ -360,9 +363,6 @@ pub struct Engine {
clock_https: Option<(f64, Instant)>,
clock_next_https: Instant,
clock_last_sample: Instant,
jobs: crate::jobrun::Jobs,
/// a remote job has the miners stopped; they restart when the runner lets go
job_hold: bool,
last_accepted: Option<Instant>,
telemetry: Option<Proc>,
telemetry_retry_at: Instant,
@ -390,6 +390,7 @@ impl Engine {
// labels never carry the hostname: two cloned PCs with the same COMPUTERNAME would sign with the same keys
let label_base = format!("{}-{}", if cfg!(target_os = "macos") { "mac" } else { "win" }, shared.runtime.id8());
let ota = crate::ota::Updater::new(&shared);
let jobs = crate::jobrun::Jobs::new(&shared);
Engine {
shared,
bins,
@ -401,7 +402,6 @@ impl Engine {
node_restart_at: None,
node_started_at: now,
node_starts: 0,
let jobs = crate::jobrun::Jobs::new(&shared);
node_restarts: 0,
node_log: None,
node_last_reading: None,
@ -420,6 +420,8 @@ impl Engine {
last_upload: now,
upload_thread: None,
ota,
jobs,
job_hold: false,
stamp,
label_base,
wrapper,
@ -431,14 +433,13 @@ impl Engine {
clock_https: None,
clock_next_https: now + Duration::from_secs(5),
clock_last_sample: now,
jobs,
job_hold: false,
last_accepted: None,
telemetry: None,
telemetry_retry_at: now,
power_busy: false,
power_restore_pending: false,
power_via_host: None,
node_override_unusable: false,
stability: std::collections::HashMap::new(),
last_stability: now,
last_settings_save: now,
@ -1321,6 +1322,7 @@ impl Engine {
self.upload_logs(false);
}
self.tick_update();
self.tick_jobs();
if now >= self.clock_next_https {
self.command(Cmd::ClockCheck);
}
@ -1332,7 +1334,6 @@ impl Engine {
if self.wrapper && now.duration_since(self.last_state_print) >= Duration::from_secs(3) {
self.last_state_print = now;
println!("STATE {}", self.shared.wrapper_state());
self.tick_jobs();
let _ = std::io::stdout().flush();
}
if now.duration_since(self.last_settings_save) >= Duration::from_secs(120) {
@ -1584,6 +1585,10 @@ impl Engine {
}
continue;
}
if self.job_hold {
// a remote job has the GPU; the miners wait until it lets go (src/jobrun.rs)
continue;
}
if paused || !synced {
if let Some(c) = self.st().mining.cards.get_mut(card_idx) {
if !paused && c.state != "failed" {
@ -1595,10 +1600,6 @@ impl Engine {
}
if self.miners[i].building {
continue;
if self.job_hold {
// a remote job has the GPU; the miners wait until it lets go (src/jobrun.rs)
continue;
}
}
if let Some(at) = self.miners[i].restart_at {
if now >= at {
@ -2057,6 +2058,7 @@ impl Engine {
fn shutdown(&mut self) {
self.shared.log("quit: stopping the miners, then the node");
self.jobs.abort(&self.shared, "the app is quitting");
self.stop_miners("quit");
self.stability_line();
if let Some(mut t) = self.telemetry.take() {
@ -2068,7 +2070,6 @@ impl Engine {
}
self.stop_node();
if let Some(mut k) = self.keep_awake.take() {
self.jobs.abort(&self.shared, "the app is quitting");
k.stop();
}
let st = self.st();

18
app/igneum-app/src/server.rs
View file

@ -235,6 +235,15 @@ fn api_post(shared: &Arc<Shared>, path: &str, body: Value) -> Result<Value, Stri
shared.send(Cmd::OpenUpdateFile);
Ok(json!({ "ok": true }))
}
"/api/jobs/allow" => {
let on = body.get("on").and_then(|v| v.as_bool()).ok_or("on missing")?;
shared.send(Cmd::JobsAllow(on));
Ok(json!({ "ok": true }))
}
"/api/jobs/check" => {
shared.send(Cmd::JobsCheck);
Ok(json!({ "ok": true }))
}
"/api/open" => {
let url = s("url").ok_or("url missing")?;
if url.starts_with("https://") || url.starts_with("http://") {
@ -263,12 +272,3 @@ fn api_post(shared: &Arc<Shared>, path: &str, body: Value) -> Result<Value, Stri
_ => Err("unknown api".into()),
}
}
"/api/jobs/allow" => {
let on = body.get("on").and_then(|v| v.as_bool()).ok_or("on missing")?;
shared.send(Cmd::JobsAllow(on));
Ok(json!({ "ok": true }))
}
"/api/jobs/check" => {
shared.send(Cmd::JobsCheck);
Ok(json!({ "ok": true }))
}

2
docs/evidence.md
View file

@ -54,7 +54,7 @@ Versions in the table: `igneum-pow` is the Rust crate at `igneum-pow/Cargo.toml`
| 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`, `8dae48b`, `6b5bd92`; fork worktree `vendor/igneum-node-harness` and `devnet-v4`; `tools/harness/`; `infra/cloud-devnet/experiments/partition.sh` | `tools/harness/` against a private `igneumd` test network; the merged node's harness scenarios 2 and 5; the cloud network's 10-minute partition of Singapore (`results/2026-10-04/partition-sin-20261004-110906/partition.md`); bench-log "consensus attack harness", "devnet-v4 integration" | 3 October 2026, Apple M5 Max: 63 malformed cases, node up on every one; withholding at 10% to 45% 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 floods left template p95 under 4 ms; one FAIL, a 45% withholder releasing every 20 blocks took 50.7% of blues (bound 47.4%). Merged node, 4 October 2026: 63 cases, node up, 0 cache builds; the 10 s timestamp floor and future bound exact. Cloud network, 4 October 2026: 12 nodes in five locations on their own chain, Singapore cut off by iptables for 10 minutes; the two minority nodes adopted the majority chain 10 and 14 s after the heal with reorgs of 445 and 516 blocks, the majority's deepest reorg was 2 blocks, 0 conflicting locks (none were possible: the weight window stood at DAA 3,030 of 7,200). CPU miners only; the finality rules under partition are row 10 | 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`, `8dae48b`; fork worktree `vendor/igneum-node-r3` branch `r3-fixes` at `5166ee26`, merged into `devnet-v4` | `measure_m15_attack_before_and_after` (ignored test, release, `--features igneum-pow`); kaspa-pow 8, header_processor 1, p2p `pow_guard` 2 tests; harness scenario 5 on the merged node | 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. Merged node, 4 October 2026: 63 harness cases with 0 cache builds (the node log shows one build, the honest day) and the M15 p2p cases disconnected by the strike guard; the live devnet v4 runs it. Measured through the validate path with `skip_proof_of_work`, not the daemon RPC | none yet |
| 29 | Blocks reach every node well inside GHOSTDAG's delay bound across continents | Litepaper Speed (GHOSTDAG at one block a second); spec 03 C1 (lock latency); `infra/cloud-devnet/README.md` | tested by the team | repo `6b5bd92`; `infra/cloud-devnet/experiments/latency.sh`, `analyze.py`; the Linux cross-build `infra/cross/build-linux.sh` | 12 `igneumd` nodes on Hetzner VMs in Helsinki, Falkenstein, Ashburn, Hillsboro and Singapore (own chain `igneum-devnet-20`, one CPU trickle miner each), a ping matrix, then 10 minutes of per-node arrival logs joined on block hash; `results/2026-10-04/latency/propagation.md` and `rtt-by-region.md` | 644 blocks in the window, 642 seen by at least 80% of nodes; arrival at a node minus the first arrival anywhere: p50 343 ms, p90 497 ms, p99 666 ms, max 2,313 ms; by region p50 239 ms (Falkenstein) to 413 ms (Singapore), p90 455 to 632 ms; inter-region RTT 35 ms (Helsinki to Falkenstein) to 289 ms (Ashburn to Singapore); first arrival minus header time median 490 ms. 4 October 2026. The network is the project's own: 12 nodes not 20 (a new account's limits), CPU hash rate only, clocks by chrony, one evening of data; the 5 s bound behind GHOSTDAG k is a design parameter this run did not challenge | none yet |
| 30 | One click: install, press start, the card mines; the app looks after its node | Homepage Mine section ("One click: install, press start"); litepaper "One click, for everyone else"; journey phase 5 | tested by the team | repo `3bb50d6`, `2c4b30f`, `6461540` (package 0.3.0: prebuilt NVRTC CUDA worker and generic OpenCL worker, driver only), `a1a33cb`, `7c794df`, `0d4498e`, `6c083db` (Igneum Miner 0.3.0), `78903cd` (0.3.1, over-the-air updates) | `Igneum-Miner-Setup-0.3.0.exe` (runner-built, unsigned) on a Windows PC with an RTX 5090 and no toolchain; `proto-cuda/nvrtc/emu/serve-check.sh` on the Mac; `proto-cuda/windows-app/TEST.md`; bench-log "one-click Windows workers", "first machine on the Igneum Miner app", "a node 60 s behind the clock is silently dead", "the gfx1036 worker fault" | One machine so far, PC 2, 4 October 2026. The NVRTC worker compiled the pack on the card with no toolchain installed and mined at 124.2 MH/s, equal to the nvcc-built worker, 0 rejected, CPU re-check clean; inside the app 117 to 119 MH/s with 34 accepted blocks in the first minute, the integrated AMD chip at 3.3 MH/s beside it (row 9). Two defects found by the install, both fixed the same hour: a clock 62 s slow after a power cut made the node reject every relayed block for 12 minutes with no visible reason (the app now reads the skew from the node's warnings, the block timestamps over the EVM RPC and an HTTPS Date header, warns over 5 s and blocks Start over 10 s, with a one-click clock sync; checked on the Mac with a fake 60 s skew; a one-line node warning is filed), and the node card said "syncing" while the miner was already accepted. The Mac could only emulate the NVIDIA path (17 of 17 sampled hashes) and the AMD path on Apple OpenCL (15 of 15). Over-the-air updates were dry-run on a private devnet (0.3.0 to 0.3.1 and back), not on a user's machine. The installer is unsigned (SmartScreen "run anyway"). Second machine, the same afternoon: a friend of the project installed Igneum Miner 0.3.1 from the DMG on an Apple silicon laptop with no toolchain and no instructions beyond five steps; the node synced from the seed, the Metal worker reported ready, 33 accepted blocks and 0 rejected in 7 minutes at 21.0 MH/s average, CPU re-check OK on every share, uploads arriving every minute under its per-install id. That laptop is not the project's hardware, but the result is observed through the project's own log intake and reported by the project, so it stays tested by the team until an outsider publishes a run of their own. The devnet's other GPU machines (PC 1 and the Mac) run the same workers through the launcher, not the app | none yet |
| 30 | One click: install, press start, the card mines; the app looks after its node | Homepage Mine section ("One click: install, press start"); litepaper "One click, for everyone else"; journey phase 5 | tested by the team | repo `3bb50d6`, `2c4b30f`, `6461540` (package 0.3.0: prebuilt NVRTC CUDA worker and generic OpenCL worker, driver only), `a1a33cb`, `7c794df`, `0d4498e`, `6c083db` (Igneum Miner 0.3.0), `78903cd` (0.3.1, over-the-air updates) | `Igneum-Miner-Setup-0.3.0.exe` (runner-built, unsigned) on a Windows PC with an RTX 5090 and no toolchain; `proto-cuda/nvrtc/emu/serve-check.sh` on the Mac; `proto-cuda/windows-app/TEST.md`; bench-log "one-click Windows workers", "first machine on the Igneum Miner app", "a node 60 s behind the clock is silently dead", "the gfx1036 worker fault" | Four machines by 15:45 BST on 4 October 2026: PC 2, then PC 1 (RTX 5090 at 110 MH/s under the 80% power cap), the project's Apple M5 Max (25 MH/s) and the outside Apple silicon laptop (row 29), all on Igneum Miner 0.3.1. The NVRTC worker compiled the pack on the card with no toolchain installed and mined at 124.2 MH/s, equal to the nvcc-built worker, 0 rejected, CPU re-check clean; inside the app 117 to 119 MH/s with 34 accepted blocks in the first minute, the integrated AMD chip at 3.3 MH/s beside it (row 9). Two defects found by the install, both fixed the same hour: a clock 62 s slow after a power cut made the node reject every relayed block for 12 minutes with no visible reason (the app now reads the skew from the node's warnings, the block timestamps over the EVM RPC and an HTTPS Date header, warns over 5 s and blocks Start over 10 s, with a one-click clock sync; checked on the Mac with a fake 60 s skew; a one-line node warning is filed), and the node card said "syncing" while the miner was already accepted. The Mac could only emulate the NVIDIA path (17 of 17 sampled hashes) and the AMD path on Apple OpenCL (15 of 15). Over-the-air updates were dry-run on a private devnet (0.3.0 to 0.3.1 and back), not on a user's machine. The installer is unsigned (SmartScreen "run anyway"). Second machine, the same afternoon: a friend of the project installed Igneum Miner 0.3.1 from the DMG on an Apple silicon laptop with no toolchain and no instructions beyond five steps; the node synced from the seed, the Metal worker reported ready, 33 accepted blocks and 0 rejected in 7 minutes at 21.0 MH/s average, CPU re-check OK on every share, uploads arriving every minute under its per-install id. That laptop is not the project's hardware, but the result is observed through the project's own log intake and reported by the project, so it stays tested by the team until an outsider publishes a run of their own. The devnet's other GPU machines (PC 1 and the Mac) run the same workers through the launcher, not the app | none yet |
## Count by status

2
site/evidence.html
View file

@ -112,7 +112,7 @@ footer{border-top:1px solid var(--line);padding-block:32px 48px;font-size:13px;c
<tr data-status="tested by the team"><td class="n">27</td><td class="claim">The node survives malformed input, floods, withholding, partitions and eclipses<div class="where">Litepaper Speed ("GHOSTDAG, the BlockDAG consensus proven on Kaspa"); spec 2</div></td><td><span class="st st-2">tested by the team</span></td><td class="mono">repo <code>394030c</code>, <code>8dae48b</code>, <code>6b5bd92</code>; fork worktree <code>vendor/igneum-node-harness</code> and <code>devnet-v4</code>; <code>tools/harness/</code>; <code>infra/cloud-devnet/experiments/partition.sh</code></td><td><code>tools/harness/</code> against a private <code>igneumd</code> test network; the merged node's harness scenarios 2 and 5; the cloud network's 10-minute partition of Singapore (<code>results/2026-10-04/partition-sin-20261004-110906/partition.md</code>); bench-log "consensus attack harness", "devnet-v4 integration"</td><td>3 October 2026, Apple M5 Max: 63 malformed cases, node up on every one; withholding at 10% to 45% 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 floods left template p95 under 4 ms; one FAIL, a 45% withholder releasing every 20 blocks took 50.7% of blues (bound 47.4%). Merged node, 4 October 2026: 63 cases, node up, 0 cache builds; the 10 s timestamp floor and future bound exact. Cloud network, 4 October 2026: 12 nodes in five locations on their own chain, Singapore cut off by iptables for 10 minutes; the two minority nodes adopted the majority chain 10 and 14 s after the heal with reorgs of 445 and 516 blocks, the majority's deepest reorg was 2 blocks, 0 conflicting locks (none were possible: the weight window stood at DAA 3,030 of 7,200). CPU miners only; the finality rules under partition are row 10</td><td class="iv">none yet</td></tr>
<tr data-status="tested by the team"><td class="n">28</td><td class="claim">Headers are validated cheaply before the lottery engine runs, so forged timestamps cannot force 256 MiB cache builds<div class="where">Spec 2.4; ledger M15</div></td><td><span class="st st-2">tested by the team</span></td><td class="mono">repo <code>0953ec7</code>, <code>8dae48b</code>; fork worktree <code>vendor/igneum-node-r3</code> branch <code>r3-fixes</code> at <code>5166ee26</code>, merged into <code>devnet-v4</code></td><td><code>measure_m15_attack_before_and_after</code> (ignored test, release, <code>--features igneum-pow</code>); kaspa-pow 8, header_processor 1, p2p <code>pow_guard</code> 2 tests; harness scenario 5 on the merged node</td><td>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. Merged node, 4 October 2026: 63 harness cases with 0 cache builds (the node log shows one build, the honest day) and the M15 p2p cases disconnected by the strike guard; the live devnet v4 runs it. Measured through the validate path with <code>skip_proof_of_work</code>, not the daemon RPC</td><td class="iv">none yet</td></tr>
<tr data-status="tested by the team"><td class="n">29</td><td class="claim">Blocks reach every node well inside GHOSTDAG's delay bound across continents<div class="where">Litepaper Speed (GHOSTDAG at one block a second); spec 03 C1 (lock latency); <code>infra/cloud-devnet/README.md</code></div></td><td><span class="st st-2">tested by the team</span></td><td class="mono">repo <code>6b5bd92</code>; <code>infra/cloud-devnet/experiments/latency.sh</code>, <code>analyze.py</code>; the Linux cross-build <code>infra/cross/build-linux.sh</code></td><td>12 <code>igneumd</code> nodes on Hetzner VMs in Helsinki, Falkenstein, Ashburn, Hillsboro and Singapore (own chain <code>igneum-devnet-20</code>, one CPU trickle miner each), a ping matrix, then 10 minutes of per-node arrival logs joined on block hash; <code>results/2026-10-04/latency/propagation.md</code> and <code>rtt-by-region.md</code></td><td>644 blocks in the window, 642 seen by at least 80% of nodes; arrival at a node minus the first arrival anywhere: p50 343 ms, p90 497 ms, p99 666 ms, max 2,313 ms; by region p50 239 ms (Falkenstein) to 413 ms (Singapore), p90 455 to 632 ms; inter-region RTT 35 ms (Helsinki to Falkenstein) to 289 ms (Ashburn to Singapore); first arrival minus header time median 490 ms. 4 October 2026. The network is the project's own: 12 nodes not 20 (a new account's limits), CPU hash rate only, clocks by chrony, one evening of data; the 5 s bound behind GHOSTDAG k is a design parameter this run did not challenge</td><td class="iv">none yet</td></tr>
<tr data-status="tested by the team"><td class="n">30</td><td class="claim">One click: install, press start, the card mines; the app looks after its node<div class="where">Homepage Mine section ("One click: install, press start"); litepaper "One click, for everyone else"; journey phase 5</div></td><td><span class="st st-2">tested by the team</span></td><td class="mono">repo <code>3bb50d6</code>, <code>2c4b30f</code>, <code>6461540</code> (package 0.3.0: prebuilt NVRTC CUDA worker and generic OpenCL worker, driver only), <code>a1a33cb</code>, <code>7c794df</code>, <code>0d4498e</code>, <code>6c083db</code> (Igneum Miner 0.3.0), <code>78903cd</code> (0.3.1, over-the-air updates)</td><td><code>Igneum-Miner-Setup-0.3.0.exe</code> (runner-built, unsigned) on a Windows PC with an RTX 5090 and no toolchain; <code>proto-cuda/nvrtc/emu/serve-check.sh</code> on the Mac; <code>proto-cuda/windows-app/TEST.md</code>; bench-log "one-click Windows workers", "first machine on the Igneum Miner app", "a node 60 s behind the clock is silently dead", "the gfx1036 worker fault"</td><td>One machine so far, PC 2, 4 October 2026. The NVRTC worker compiled the pack on the card with no toolchain installed and mined at 124.2 MH/s, equal to the nvcc-built worker, 0 rejected, CPU re-check clean; inside the app 117 to 119 MH/s with 34 accepted blocks in the first minute, the integrated AMD chip at 3.3 MH/s beside it (row 9). Two defects found by the install, both fixed the same hour: a clock 62 s slow after a power cut made the node reject every relayed block for 12 minutes with no visible reason (the app now reads the skew from the node's warnings, the block timestamps over the EVM RPC and an HTTPS Date header, warns over 5 s and blocks Start over 10 s, with a one-click clock sync; checked on the Mac with a fake 60 s skew; a one-line node warning is filed), and the node card said "syncing" while the miner was already accepted. The Mac could only emulate the NVIDIA path (17 of 17 sampled hashes) and the AMD path on Apple OpenCL (15 of 15). Over-the-air updates were dry-run on a private devnet (0.3.0 to 0.3.1 and back), not on a user's machine. The installer is unsigned (SmartScreen "run anyway"). Second machine, the same afternoon: a friend of the project installed Igneum Miner 0.3.1 from the DMG on an Apple silicon laptop with no toolchain and no instructions beyond five steps; the node synced from the seed, the Metal worker reported ready, 33 accepted blocks and 0 rejected in 7 minutes at 21.0 MH/s average, CPU re-check OK on every share, uploads arriving every minute under its per-install id. That laptop is not the project's hardware, but the result is observed through the project's own log intake and reported by the project, so it stays tested by the team until an outsider publishes a run of their own. The devnet's other GPU machines (PC 1 and the Mac) run the same workers through the launcher, not the app</td><td class="iv">none yet</td></tr>
<tr data-status="tested by the team"><td class="n">30</td><td class="claim">One click: install, press start, the card mines; the app looks after its node<div class="where">Homepage Mine section ("One click: install, press start"); litepaper "One click, for everyone else"; journey phase 5</div></td><td><span class="st st-2">tested by the team</span></td><td class="mono">repo <code>3bb50d6</code>, <code>2c4b30f</code>, <code>6461540</code> (package 0.3.0: prebuilt NVRTC CUDA worker and generic OpenCL worker, driver only), <code>a1a33cb</code>, <code>7c794df</code>, <code>0d4498e</code>, <code>6c083db</code> (Igneum Miner 0.3.0), <code>78903cd</code> (0.3.1, over-the-air updates)</td><td><code>Igneum-Miner-Setup-0.3.0.exe</code> (runner-built, unsigned) on a Windows PC with an RTX 5090 and no toolchain; <code>proto-cuda/nvrtc/emu/serve-check.sh</code> on the Mac; <code>proto-cuda/windows-app/TEST.md</code>; bench-log "one-click Windows workers", "first machine on the Igneum Miner app", "a node 60 s behind the clock is silently dead", "the gfx1036 worker fault"</td><td>Four machines by 15:45 BST on 4 October 2026: PC 2, then PC 1 (RTX 5090 at 110 MH/s under the 80% power cap), the project's Apple M5 Max (25 MH/s) and the outside Apple silicon laptop (row 29), all on Igneum Miner 0.3.1. The NVRTC worker compiled the pack on the card with no toolchain installed and mined at 124.2 MH/s, equal to the nvcc-built worker, 0 rejected, CPU re-check clean; inside the app 117 to 119 MH/s with 34 accepted blocks in the first minute, the integrated AMD chip at 3.3 MH/s beside it (row 9). Two defects found by the install, both fixed the same hour: a clock 62 s slow after a power cut made the node reject every relayed block for 12 minutes with no visible reason (the app now reads the skew from the node's warnings, the block timestamps over the EVM RPC and an HTTPS Date header, warns over 5 s and blocks Start over 10 s, with a one-click clock sync; checked on the Mac with a fake 60 s skew; a one-line node warning is filed), and the node card said "syncing" while the miner was already accepted. The Mac could only emulate the NVIDIA path (17 of 17 sampled hashes) and the AMD path on Apple OpenCL (15 of 15). Over-the-air updates were dry-run on a private devnet (0.3.0 to 0.3.1 and back), not on a user's machine. The installer is unsigned (SmartScreen "run anyway"). Second machine, the same afternoon: a friend of the project installed Igneum Miner 0.3.1 from the DMG on an Apple silicon laptop with no toolchain and no instructions beyond five steps; the node synced from the seed, the Metal worker reported ready, 33 accepted blocks and 0 rejected in 7 minutes at 21.0 MH/s average, CPU re-check OK on every share, uploads arriving every minute under its per-install id. That laptop is not the project's hardware, but the result is observed through the project's own log intake and reported by the project, so it stays tested by the team until an outsider publishes a run of their own. The devnet's other GPU machines (PC 1 and the Mac) run the same workers through the launcher, not the app</td><td class="iv">none yet</td></tr>
</tbody></table></div>
<p class="note">Click a column heading to sort; click again to reverse. Versions: <code>igneum-pow</code> is the Rust crate at version 0.2.0 (generator version 2, 4 October 2026); repo commits are this repository's; fork commits are the node fork and its worktrees, named by message as the engineering log names them. Source of every number: the <a href="/bench">engineering log</a>. The source of this page is <code>docs/evidence.md</code> in the repository.</p>
<h2>What would move a row</h2>