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

attack-pass: F8 FIXED-AND-PASSED in class v4 sub-version 3 (c3d32437: 60 of 64 under 1.2x, the named four-seed tail, 0 exhausted); the pass record reads PASS or FIXED-AND-PASSED on every row

Browse source 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs 2026-10-07 16:03:18 +00:00
parent c29e55bb76
commit d878ee830d
1 changed files with 26 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

33
docs/analysis/attack-pass-2026-10.md
View file

@ -13,8 +13,14 @@ Lane: attack-pass, worktree `igneum-wt-attack`, branch `attack-pass` from `origi
Binary built on igneum-build-1 (ELF x86-64, `igneum-pow` 0.2.0, sha256 6d2867...1a9ebe5) and run there under
the box's slots; model and era work from `sim/horizon/algorithm/model.py` and `infra/fast-time/`. Each row below
carries the method, the known-failed shape where one exists, the result with numbers, and PASS, RUNNING,
BLOCKED or FINDING. The freeze tag waits on every row reading PASS or FIXED-AND-PASSED; the era VDF (F7 sub-row a) landed in the node on
7 October 2026 with the re-roll harness silent against it (0 of 6 cuts), so that precondition is met. Main checks every number
BLOCKED or FINDING. PASS RECORD (7 October 2026, 16:0x UTC, 17:0x UK): every row reads PASS or FIXED-AND-PASSED. F1 PASS (AP-F1-1 on the
v5 list at 3.0 percent); F2 PASS, effort-bounded; F3 PASS; F4 PASS against class v4 (AP-F4-1 on the v5 list); F5
FIXED-AND-PASSED (the F2 hour skipped by decision); F6 PASS (the worst of 10^5 programs 8.708 ms on the half-core
proxy; O-1.14 closed on an i7-9700K); F7 PASS on all three sub-rows (the era VDF in the node, 0 of 6 re-rolls); F8
FIXED-AND-PASSED in class v4 sub-version 3 at 017e7037 (AP-F8-1, AP-F8-2, AP-F8-3 ours and closed; the four-seed
unattributed tail named); F9 PASS on (a) and (c), (b) closed by the same fix; F10 PASS. Frozen generator: class v4
sub-version 3, igneum-pow 017e70376489251e18564c0abce7e466e606c8b3, devnet epoch-0 id a785001687d8688a. The
freeze tag `cryptanalysis-target-1` is the coordinator's cut on this record; the era VDF precondition is met (F7 a). Main checks every number
against the log before quoting it to the project lead.
## Status board
@ -28,8 +34,8 @@ against the log before quoting it to the project lead.
| F5 | Chip-model sweep + AWS F2 FPGA hour | evidence row 17 holds across the sweep; FPGA row under 27 M reads/s/W | sweep: 2.1x at k=1 GDDR7 reproduces, 3.2x at k=0.5, 4.1x at k=0.3 (matches ledger M32); FPGA row 2.3 to 2.9 G/s, 10 to 20 M reads/s/W (literature). FINDING: the k=0.33 figure is framed as the X9's measured core (M32) and a "measured class" (ladder branch §5a); the X9 was withdrawn before launch and never benchmarked. F2 hour SKIPPED: no AWS account | FIXED-AND-PASSED (sweep PASS; AP-F5-1 fixed and re-gated 7 Oct 2026: chip section re-run 2.1x at k=1 unchanged, identity grep 0 hits, site lane concurred); F2 hour SKIPPED-BY-DECISION (the project lead, 7 Oct 2026, 09:5x UK; plan 4.2 row F5 is the sweep only at 3714c2a0; the FPGA row stays the JEDEC-ceiling model row labelled unmeasured) |
| F6 | Verifier worst case over 10^5 programs + O-1.14 laptop run | worst program under 10 ms cold on the half-core proxy and the laptop | 100,000 programs ranked, 50,000 timed cold one-core (worst 6.194 ms), the worst 200 re-timed and the worst 1,000 timed on the half-core proxy under the per-core lease (core 40 at 3,799.9 MHz): worst 8.708 ms (`attack-f6/87142`), 1.29 ms under the gate, every half-core reading under 9 ms; dr736 fails as it must (15.49). O-1.14 CLOSED on an i7-9700K (v4 6.334 ms cold max, dr736 10.04 fails). Ladder ceiling from the worst program on the half-core proxy: N about 300,000, so rung 2 admissible, rung 3 not. Record `docs/analysis/attack-pass/f6-verifier.md` | PASS |
| F7 | Era-draw bias harness + 2^20 era-seed census | no re-roll inside the publish window; no era class with gain over 1.1x over 2^-20 | (b) census PASS at 2^24 seeds (no class over 1.1x, stride bijective, R, pos and M uniform, planted cases fire); (c) 64-bit day-key seeding PASS (0 collisions in 2^17); (a) PASS with the era VDF in the node (era-vdf lane, fork era-vdf-node 394a5902 on release-0.3.20-node c4459193, behind era_vdf_activation_daa; master a4eaf766 carries the spec text, `docs/analysis/era-vdf-2026-10-07.md` and `tools/era-vdf/reroll.mjs`): against the real era cut on three fast-time nodes the re-roll harness fires with the VDF off (6 of 6 cuts) and is silent with it on (0 of 6; the adversary's 5.4 to 6.6 s evaluation against a 1 s block interval, the honest chain 3 to 10 blocks ahead, three nodes agreeing on every era seed); the delay is 517 s on the fastest prover measured (chiavdf NUDUPL over GMP, 208.8K squarings/s) at T = 108,000,000, 259x the 2 s window. Open, not a gate: the verify is 22 ms against the 10 ms target (O-4.6, 0.3.22). Record `docs/analysis/attack-pass/f7-era.md` | PASS (a, b, c) |
| F8 | Uniformity censuses (line-index 2^28, distinct lines, cross-hash histogram) | uniform within the window model of spec 1.13.1, layer 8; the excess beyond it within 6 sigma over 64 seeds; no hot set under 1% of items beyond the model | line index PASS at 2^28. Phase E (6 Oct stream): 31 of 64 seeds over 1.2x, the lossy load-source class, AP-F8-1. Re-gate on sub-version 1 (8c728ca3, 0.3.20): 11 of 64 over 1.2x, worst 29.27x (p31), a constant delivered through an admitted writer or across the iteration boundary; three residual classes named; sub-version 2 (freshness fixpoint + dynamic source count) is the fix. F9's harness retired from the re-gate (counts the windows) | FINDING; sub-version 1 FAILS 11 of 64, sub-version 2 FAILS 9 of 64 (worst 4.82x; one low-entropy site per program, lineage-blind, the acceptance never ran the shadow: AP-F8-3); AP-F8-2 exhaustion FIXED-AND-PASSED at 8bdcbdd8; sub-version 3 is the target |
| F9 | Acceptance edges (39) + header grinding on an RTX 5090 | zero passing programs with a hot set under 1%; grinding gain under 1% of rate | (a) edges on generator 4 over 10^5 seeds: 34 disagreements in 105,064 candidates (29 const_bit, 4 bias, 1 lane_const), every one a one-bit or sampling-noise property moving the choice to an attempt both stand-ins accept, nothing in the attacker's favour: PASS; (b) hot-set search over 10^6 seeds: 11,696 passing programs (1.17%) concentrate 1% or more of reads on a hot set, worst 17.3%: FINDING, the same or-saturation load-source class as AP-F8-1 found by a second harness (F9-1 merged into AP-F8-1), re-gated on the amended stream; (c) grinding on the 5090: +0.004% at K = 2^14, ceiling +43%: PASS. Record `docs/analysis/attack-pass/f9-grind.md` | (a) PASS; (b) FINDING = AP-F8-1 class (10^6 seeds on the 6 Oct stream); the harness is retired from the re-gate (its metric counts the era's windows), F8's census re-gates; (c) PASS |
| F8 | Uniformity censuses (line-index 2^28, distinct lines, cross-hash histogram) | uniform within the window model of spec 1.13.1, layer 8; the excess beyond it within 6 sigma over 64 seeds; no hot set under 1% of items beyond the model | line index PASS at 2^28. The 6 Oct stream: 31 of 64 seeds over 1.2x (AP-F8-1, the lossy load source). Sub-version 1 (8c728ca3): 11 of 64. Sub-version 2 (07a809a7 / 8bdcbdd8): 9 of 64; AP-F8-2 (exhaustion) closed there, 0 of 10^6. Sub-version 3 (017e7037: the acceptance executing the shadow block, AP-F8-3; the shared-operand rule; (c'') the distinct-index ratio): 60 of 64 under 1.2x, the four over the named unattributed tail at 1.22x to 1.50x with no chip consequence; 0 exhausted in 24,631 chain-shaped seeds, the draw total by construction. Record `docs/analysis/attack-pass/f8-uniform.md` | FIXED-AND-PASSED (sub-version 3, 017e7037, the frozen generator) |
| F9 | Acceptance edges (39) + header grinding on an RTX 5090 | zero passing programs with a hot set under 1%; grinding gain under 1% of rate | (a) edges on generator 4 over 10^5 seeds: 34 disagreements in 105,064 candidates (29 const_bit, 4 bias, 1 lane_const), every one a one-bit or sampling-noise property moving the choice to an attempt both stand-ins accept, nothing in the attacker's favour: PASS; (b) hot-set search over 10^6 seeds: 11,696 passing programs (1.17%) concentrate 1% or more of reads on a hot set, worst 17.3%: FINDING, the same or-saturation load-source class as AP-F8-1 found by a second harness (F9-1 merged into AP-F8-1), re-gated on the amended stream; (c) grinding on the 5090: +0.004% at K = 2^14, ceiling +43%: PASS. Record `docs/analysis/attack-pass/f9-grind.md` | (a) PASS; (b) the AP-F8-1 class on the 6 Oct stream, closed by sub-version 3 (F8's census is the re-gate instrument; this harness's hot-share metric counts the era's windows); (c) PASS |
| F10 | Ladder signal monotonicity harness | no step without 90% over 7 windows in either direction | known-fail fails, known pass passes; new cases on the exact-share driver: 89% up holds (no step), 89 then 90% down with restarts steps only at 90% after the 7-window cool-down, the floor holds under 100% down (never below rung 0); decision per seed block memoised, identical after a restart, never differs between nodes; 19 of 19 checks per case. Two items to main, not findings: a stale commit string in the ladder lane's igneumd, and proof-synced nodes deciding rung 0 until the witness lands (a precondition line for spec 01). Record `docs/analysis/attack-pass/f10-ladder.md` | PASS |
## The rows
@ -496,9 +502,22 @@ share their value with two clean seeds (p44 0.9612, p52 0.9613), so the 2^24 sta
p34 stay the open tail, unattributed. The ratio refuses about 4 percent of candidates that pass every other test
(4,099-program census at 017e7037: mean attempts 2.086 against 1.998, max 17, 0 lossy-sourced load sites of 65,584,
0 exhaustions; suite 103 of 103; devnet epoch-0 at attempt 1, id a785001687d8688a, pairing verified by this lane).
Status: FINDING-OPEN; sub-version 1 FAILS the re-gate (11 of 64); sub-version 2 FAILS it (9 of 64, worst 4.82x);
FIXED-AND-PASSED on the 64-seed verdict against sub-version 3 (the acceptance executing the shadow block, the
per-site distinct-index ratio rule, the cap and the last resort).
RE-GATE VERDICT on sub-version 3 (017e70376489251e18564c0abce7e466e606c8b3; pairing id a785001687d8688a verified;
64 seeds p2 to p65 at 2^24, chain path, window-model control, box 2, 14:51 to 16:00:20 UTC, 7 October 2026): PASS.
60 of 64 under 1.2x (0.9915x to 1.144x); the four over are the named open tail, unattributed and chased: p10
1.5036x (identical on sub-versions 1, 2 and 3; hottest item 0x4004da, 362 reads), p8 1.3776x (0x837de4, 420), p34
1.2505x (0x800010, 541, the one-bit value through sub at 5), p4 1.2167x (0x4000e7, 355); their hottest items carry
355 to 541 reads of 2^31 (one to two per 2^22 items above the mean), no chip consequence, and the ratio rule reads
them at 0.9927 to 0.9963 at 2^20, inside the clean spread. Every strong seed of sub-versions 1 and 2 is under the
line (p23 4.82x to under 1.2x, p19, p15, p18, p56 likewise). Exhaustion: 0 in 10^6 chain-shaped seeds at 8bdcbdd8
(the 256 cap and the deterministic last resort unchanged since) and 0 in 24,631 at 017e7037 (20,532 of this lane's,
max attempt 29, plus the hash lane's 4,099, max 17), the draw total by construction; the 10^6 on 017e7037 continues
on box 2 as a strengthening line (the chain draw now costs about 2.2 s per candidate through (c''), so about two
days) and is not a condition. Node consequence, not a gate: about 2 attempts at 2.2 s each per epoch per node, 4 to
5 s at one epoch an hour. Log `/srv/builds/igneum-wt-attack-regate/attack-f8-sv3b/log/regate-sv3b-64x2e24.log`.
Status: FIXED-AND-PASSED. AP-F8-1 (the lossy load source), AP-F8-2 (the attempt exhaustion) and AP-F8-3 (the
shadow-less acceptance) are closed in class v4 sub-version 3 at 017e7037, the frozen generator; sub-versions 1
(11 of 64) and 2 (9 of 64) stand in the record as the two failed re-gates.
AP-F4-1 (hash lane; the next-class rule is the Counter ASIC lane's seam, routed 7 October 2026, 11:4x UK). A bound
on the day-key draw, not a weak class: on the M1 metric (every multiply in LUT adders, adders per mixer application