Conflict resolved: publish-jobs.sh keeps master's verify command and --tries (the retrying live check) alongside the
build kind's arguments; the usage range covers the merged header.
Conflicts resolved: state.rs keeps both the sweep fields (miner-eff) and the race fields (miner-perf); bench-log.md
keeps both appended entries; publish-manifest.sh keeps master's --override implementation (da4d413, the "every
height switch" rule, --verify-only, --tries, the retrying live check) and adds miner-perf's --tuning / --no-tuning
with the carry-over of consensus.override and tuning from the current manifest. One --override case, one parser.
The first Windows update over the air (0.3.3 to 0.3.4) sat on 'the installer is starting' with no helper log on
PC 2. The helper launched by a remote job with the same arguments ran at once. Fix in 0.3.5; the Windows machines
take 0.3.5 by hand once more.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Header: site/partials/nav.html, head.html, footer.html, injected by site/build.mjs between markers in every page (the bench template too) with the active link marked; the master mark in its black square, the wordmark, Litepaper, Live devnet, Engineering log, Evidence, GitHub and the miner button, in that order everywhere; one menu script (closes on a link, a tap outside, Escape with focus back on the button, a resize past 900 px); a skip link; one focus ring and one reduced-motion rule.
Litepaper: the hash routed by a click delegate and popstate instead of the browser's fragment jump. Root cause of the tabs fault: show() scrolled to the top of the article, never to the section; in whole-paper mode the fragment jump landed on the section and the smooth scroll then dragged the reader back to the Abstract; in one-section mode a deep link scrolled while the section was still display:none and landed on the cover. Now: every section and all 29 subsection headings reachable by URL in both modes, scroll-margin from the measured bar height (nav, plus the contents row on a phone), focus moved to the heading, a re-aim after a late font batch, mode buttons stacked in the sidebar.
Pages: homepage headline capped at 9cqw so it is two lines at 390 and 1440 (was four with "fire." alone), journey inlined by the build (no fetch, no shift), dates as "4 Oct 2026"; live page legend no longer forces 420 px of width at 390, the hash-rate unit as small mono, idle redraw at 30/s with a 1.5x backing store (0.7 to 1.2% of one core idle in Chrome for Testing, taken under load), the proving line wraps on a phone; evidence sort headings are buttons, chips carry aria-pressed, a scroll hint under 1,140 px; bench contents in a sticky scroll box, og.png?v=3; 404.html; sitemap with /evidence; vercel.json caches fonts a year and images a day.
Fonts: eight latin woff2 files in site/fonts (139 KB for all, 118 KB a typical page), preloaded first-paint faces, fallback faces with size-adjust and ascent overrides from the font tables; every page lost its render-blocking fonts.googleapis.com stylesheet and the gstatic origin.
HTML per page before to after (gzip): / 17,265 to 21,810 B; /litepaper 20,556 to 24,957; /live 15,640 to 17,777; /evidence 18,899 to 22,493; /bench 94,723 to 98,289. Third-party requests before to after: 3 to 1 (jsdelivr, light client), 1 to 0, 3 to 0, 1 to 0, 1 to 0. Lighthouse before and after is queued under the measure lock (held by reliability runs since 19:16Z, load average 258) and recorded in docs/design/site-polish-2026-10-04.md when it runs.
Checks: node site/build.mjs, tools/ci/link-check.mjs (246 links, 0 broken), identity grep, and a Chrome for Testing run over every page at 390, 768 and 1440 (no overflow, no console errors, menu, router in both modes, deep links on reload, keyboard).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
src/sweep.rs (new): the cap steps 100% to 50% in 10% steps clamped to the card's limits, per-step rows (mean draw
from nvidia-smi power.draw, mean worker interval rate), the choice (best MH/W, ties to the higher rate then the lower
cap), the nvidia-smi power parser, a state machine on an explicit clock (15 s settle, 60 s hold, 30 s cap readback
limit), the elevated helper scripts (one administrator prompt per sweep: a command file polled by one elevated
process, self-restoring after 20 idle minutes), the unsupported reasons (Apple silicon, AMD). 9 unit tests with
PC 1's recorded RTX 5090 numbers (575 W default, 460 W cap, 290 W draw, memory temperature [N/A]).
Engine: scheduler (once after install, then weekly; one card at a time; only while the card mines, after 120 s
steady, never under a remote job hold, a pause, or inside 600 s of the hour boundary), the cap-mode probe (direct
when the engine runs elevated, else the helper), abort on any fault (card leaves mining, worker error, GPU 90 C,
job, pause, quit) with the cap restored, the chosen cap held and recorded, SWEEP table lines in the app log,
--sweep mode (sweep every supported card, print the table on stdout, leave the caps, quit). Cap floor 50% (was 60).
A readback that matches the asked cap now counts as applied (PC 1 showed "cap NOT applied" for hours at 460 W).
Dashboard: live eff MH/W on each tile, the sweep line (phase, last result, or why unsupported), Sweep now / Stop /
Unpin, "pinned" and "chosen by the sweep" on the cap line, the Settings toggle, the cards-page note, slider min 50.
A cap moved by hand pins the card: the sweep records but does not change it.
PC 1 measurement: relay/playbooks/sweep-5090.ps1 (a run job, elevated, miners stopped; a second engine with --sweep
in a scratch data folder, RESULT SWEEP lines) and docs/plans/miner-eff.md with the publish command. Not published.
Untested on a card.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Under emulation (or --race with no other name) the race no longer times base alone: emu/test.sh passes again
(9 source checks, the serve protocol with prepare, swap and self-heal, 17 sampled hashes equal to igneum-pow).
Mac table: g256 (256 threads per threadgroup) +17.3% and +21.2% over the shipped 32-thread groups on two
programs, with the live app's worker sharing the GPU; the serve check found the unfair mutex (fixed in 556d057).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
New job kind `build` (jobs.rs, jobbuild.rs, jobrun.rs run_build): free-space check on both sides (20 GB), the
build-inputs zip by sha256, setup inside the distro as root (mingw-w64 posix, clang for bindgen, protoc, zstd, the
Windows rust target; idempotent), sources extracted with the target dir persisting under /root/igneum-build,
cargo build --release native and for x86_64-pc-windows-gnu, cargo test for the manifest's packages, binaries
zstd-compressed and sent to the relay (fn=upload, Blob PUT, fn=drop; 50 MB each) with sha256 in RESULT lines,
STAGE lines with UTC times, a 40-minute default budget and per-stage caps, the Linux side killed on a cap. The
app's runner stays serial (one Active at a time), so a build never overlaps a shard job; nothing stops the miners.
From this version an unknown job kind is skipped by the app (parse_lenient) instead of rejecting the whole file;
the signer stays strict.
Mac side: packaging/windows/push-build-inputs.sh packs a fork worktree, app/igneum-app, brand/icons and
proto-cuda with a manifest (branch, commit, dirty, builds, tests) and the sha256; publish-jobs.sh add --kind build;
tools/build-job.mjs packs, publishes, watches, fetches, checks both sha256 per file and the PE header of every exe
(plus verify-exe.py on igneum-app.exe), and places the binaries where push-inputs.sh, make-payload.sh and the
cloud-devnet scripts look. relay.mjs drop <file> --body carries the body.
Tested on the Mac: 33 app tests (6 new) and the signer's 21; cargo check for x86_64-pc-windows-gnu; the packer
(7.9 MB zip, no target dirs); the publisher against a scratch folder with the rebuilt signer, the old signer
refusing the kind, a bad job refused at signing; the fetch path against the live relay with a real exe (sha256
and PE pass, a wrong sha256 refused; test items deleted). Not run on a PC: the job itself. docs/plans/build-job.md
has the first job for PC 1 and the rollout order (0.3.4 must be on the PCs before a build job is published).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The engine turns a worker's race line into one TUNING {json} line in the app log (card model as the worker names it,
driver, arch, program class loads and wide loads, every variant's MH/s, winner, gain, the card's power cap and
draw, MH per watt), which the existing intake receives; the card state carries the variant for the dashboard and
one event per race. tools/tuning.mjs aggregates the records from miner_logs per card model (median MH/s or MH per
watt, at least 3 samples, de-duplicated per race) and writes tuning.json; publish-manifest.sh --tuning puts it in
the signed manifest (and now takes --override for consensus.override; both are carried over from the current
manifest when not given, --no-tuning drops it); manifest.rs parses it; ota.rs writes <app data>/tuning.json and
removes it when the manifest drops it; procs::spawn takes an environment and every miner starts with
IGNEUM_TUNING_FILE, which its worker reads at every prepare. Dry run of the publisher against a scratch folder:
tuning and override written, carried over, dropped, signature verified.
docs/plans/miner-perf.md: the signed jobs for PC 1 (fetch the race build of the NVRTC worker, then
relay/playbooks/race-5090.ps1 with the miners stopped: 17 variants, 3 rounds, twice) with the exact publish
commands for the main session; not published by the agent.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
the project lead, 4 Oct 2026 evening: "we need to make our miner better than anything else can be". The compile-ahead pipeline
built one kernel per program; it now builds a catalogue (unroll 2 or 8; the dataset load path __ldg, __ldcg, __ldcs
on NVIDIA; a register budget by --maxrregcount or __launch_bounds__, max_total_threads_per_threadgroup on Apple;
2, 4, 8 warps per block or 64 to 256 threads per threadgroup; combinations), self-tests each against the pack's
vector warps (NVIDIA) or the base kernel over 2^16 nonces (Metal), bit for bit or out, and times each for about two
seconds with the job loop paused (one mutex, mining resumes between variants). Base is the pack's text as shipped,
always first, never discarded; a race has a budget (default 120 s against the 600-DAA lead) and keeps the best so
far when it runs out, so the swap is never delayed. One line per race: variants, MH/s each, winner, gain, time.
A tuning file (--tuning, IGNEUM_TUNING_FILE) pins a variant or orders the candidates per card model.
NVIDIA: textual rewrites on the pack's own kernel_bound.cu with exact anchors from igneum-pow's emitter, so the
pack format, the miner and igneum-pow are untouched and old packs race. --race --pack <dir> runs the race alone.
Under IGNEUM_EMU the race is off (the stand-in checks the handed-over text is the pack's). Metal: the MSL hooks in
generateMSL, a lock-protected kernel slot per program, a pair compiled inline races after its first job,
--race-test --seed --day runs the race alone with a table. OpenCL is not raced yet. docs/design/miner-tuning.md.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The intake key sits in every miner package, so the relay now lets it report only (drop text and files, ack, done,
register, upload). Posting a run or task, or renaming and re-roling a machine, needs the console token.
The prove host wrote proofs through SP1's unbuffered save: on WSL2 under /mnt/c the 18 MB core proof of a shard
took longer to save than to prove. Proofs now go through a 4 MB buffer with a timed 'saved' line, and
prove-shard.sh keeps results on the Linux side and copies them per stage. Ledger P20 updated.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Observer (tools/observer/observer.mjs): reads the execution layer's JSON-RPC of a node on the proving build
(IGNEUM_EVM_RPC, default the Mac app's node 26800): every chain block's shard plan as it joins the chain
(igneum_getShardPlan by blockHash, one live_proofs row per shard, planned), the proof records of the chain
blocks of the last 10 minutes polled in rotation (igneum_getProofRecords, four in flight, 40 blocks per tick
while active, 10 before activation): proving (in the pool), verified (SP1 proof verified, or carried and checked
by consensus), paid (a carrying segment paid it), with the prover's id8, the carrier, lag in DAA and the payout.
live_state.proving = {supported, active, activation_daa, tip_daa, verifier, pool, blocks_10m,
blocks_fully_proven_10m, shards_proven_10m, shards_paid_10m, median_proof_lag_s, provers_10m}. A node without
the RPCs gives supported false (rechecked every 5 min); an unreachable endpoint is retried every 20 s. Events:
proving (activation, first paid shard), prover_seen. Additive schema (live_proofs, live_state.proving).
API (site/api/live.mjs): proving, and per block shards: [{i, n, state, prover, lag, payout (IGN), pgas}] and
proven; ?window=N (30 to 300 s) for the page's diagnostic long view; LIVE_TABLE_PREFIX reads a test observer's
tables.
Live page (site/live.html), the design change of 4 Oct 2026: three thin strips sharing one time axis, newest at
the right. BLOCKS keeps the per-miner lanes, chain path, blue/red/pending colouring, arrival glow and tooltips;
the lock ring, dashed lock line and final band leave it. FINALITY is an 18 px bar: ember wash = final (up to the
newest locked checkpoint on screen), molten tick = locked checkpoint, faint = proposed, one label at the newest
lock ("locked #522, 12 s ago"); while finality is not active it reads "finality paused: N% of weight silent" and
nothing else (R4.6.3). PROVING shows one cell per shard under each chain block, outline (planned), molten
(proving), prover colour (verified), tick (paid), a dashed "proofs land N s behind the tip" line, or the one
honest line before activation ("Proving layer: not yet activated on this devnet; activation at DAA N" / "node
without proving"). Header stats: on screen, chain, identities, last lock, proven. Legend: one line per strip.
Hover and tap tooltips on blocks and cells (block, shard, prover, lag, payout). Lanes snap on resize (they used
to ease from a zero-height layout). Phone width, no horizontal scroll; draw 0.6 ms avg, 1 ms max with 110 blocks
on screen (playwright, 1280 px).
Hero (site/index.html): a faint second glint behind a real block once every shard of it is verified, only while
the proving layer is active; pace and sampling untouched.
Verified on the private 3-node proving network (tools/proving-v0/run.mjs --network-only, activation 60) with a
CPU prover loop signing as v0/v1/v2: records relayed, verified on node 0, carried and paid (block 155 by 405,
lag 259 DAA, 0.634 IGN); screenshots in docs/design/live-proving (devnet before activation at 1280 and 375 px,
test network active, the ?window=300 view with paid cells). The live devnet shows the "not yet activated;
activation not set" line once the observer runs this build.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
rollout-v2.sh stages the Linux igneumd (gateways from the Mac, private nodes from their gateway), rolls one node at a time with
difficulty_v2_activation_daa in every override file, checks the common chain and watches the height; results/2026-10-04/
rollout-v2*.log and v2/ (the hash-rate step under v2 and the v1 comparison). docs/plans/difficulty-v2-rollout-devnet.md: the
binaries and their sha256, the activation rule (N = DAA at publish + 10,800; baked at the cut as DAA + 14,400), the exact
restart lines for the observer node, the seed and Mac node 1, the OTA path for the two PCs through NODE_OVERRIDE_PARAMS in
packaged-config.sh (the engine side landed in 00fff5f), the rehearsal record. Bench-log entry.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
partition.sh adapted to private-network mode: the cut sits on the region's gateway (INPUT, OUTPUT and FORWARD
against the far gateways' public IPs over 26611 and the DNAT ports 27001:27099), since a per-node port-26611 rule
leaves the DNAT links up. It now records the locks per side at cut, during, at heal and after convergence, the
first lock after the heal from the journals, and the heal time as each minority node's first chain removal of 5+
blocks (the sink-count criterion is tip churn on a healthy network). hop.sh and partition.sh hold the Mac awake
with caffeinate; analyze.py gains a 10-s hop series (difficulty, block count, 1- and 2-min rates, threads) and an
overshoot table; collect.sh writes hop-series.tsv and hop.md and gzips the journals.
Results 2026-10-04: partition 1 (window still filling) reorg 431/496 on the minority, 2 on the majority, healed in
10 and 14 s; partition 2 (locks active): minority locked nothing during the cut, majority locked every interval at
66.8% to 84.5% of total, 0 conflicting locks over 107 indices, healed in 11 and 15 s, first lock after heal 13 s.
Hash-rate steps x1.42, x0.70, x0.75, x1.32: difficulty overshoots x1.67, x0.66, x0.53, x1.60, settle 751 s,
never in 900 s, 241 s, 646 s. Failures stated in summary.md: the Mac hibernated during the hop (phase 2 ran 94
min), the first partition could not see locks, the script's heal and first-lock figures were artefacts (fixed),
and another agent's v2 rollout restarted every node during the second heal.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
