adv-accept report: row 90 result (10,000 seeds), the four deepest seeds against the 0.995 floor, six false refusals of nine; logs and rows copied
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs
parent
43eaab7ac7
commit
c06bed98fb
4 changed files with 10174 additions and 5 deletions
|
|
@ -0,0 +1,34 @@
|
|||
lease: holding 32 pool cores (49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80, waited 1 s, class adv): adv-accept attempts-10k-a
|
||||
adv-accept attempts: seeds 2..10002 (10000), 32 threads, cap 256, force-exhaust false
|
||||
progress: 500 of 10000 seeds in 41 s
|
||||
progress: 1000 of 10000 seeds in 81 s
|
||||
progress: 1500 of 10000 seeds in 128 s
|
||||
progress: 2000 of 10000 seeds in 176 s
|
||||
progress: 2500 of 10000 seeds in 223 s
|
||||
progress: 3000 of 10000 seeds in 270 s
|
||||
progress: 3500 of 10000 seeds in 319 s
|
||||
progress: 4000 of 10000 seeds in 367 s
|
||||
progress: 4500 of 10000 seeds in 416 s
|
||||
progress: 5000 of 10000 seeds in 461 s
|
||||
progress: 5500 of 10000 seeds in 512 s
|
||||
progress: 6000 of 10000 seeds in 562 s
|
||||
progress: 6500 of 10000 seeds in 608 s
|
||||
progress: 7000 of 10000 seeds in 655 s
|
||||
progress: 7500 of 10000 seeds in 700 s
|
||||
progress: 8000 of 10000 seeds in 745 s
|
||||
progress: 8500 of 10000 seeds in 794 s
|
||||
progress: 9000 of 10000 seeds in 841 s
|
||||
progress: 9500 of 10000 seeds in 894 s
|
||||
progress: 10000 of 10000 seeds in 940 s
|
||||
attempts: 10000 seeds in 940 s; 21119 rejected candidates, 10000 accepted, 0 exhausted (cap 256)
|
||||
rejection by part over all candidates: (a) stale=2463 (11.7%) (b) no-inject=658 (3.1%) (a') unfresh=17585 (83.3%) (c) const-bit=86 (0.4%) (c) lane-const=0 (0.0%) (c) saturated=68 (0.3%) (c') sat-source=13 (0.1%) (c'') repeated=0 (0.0%) (c'') low-entropy=238 (1.1%) (c) bias=0 (0.0%) (c) distinct=8 (0.0%)
|
||||
accepted-attempt histogram 0..24: [3227, 2201, 1435, 981, 732, 449, 308, 211, 137, 118, 64, 41, 34, 16, 12, 10, 9, 5, 3, 4, 1, 1, 0, 0, 1]
|
||||
attempt 0: 10000 candidates, rejected 6773 (67.7%): (a) stale=790 (b) no-inject=199 (a') unfresh=5667 (c) const-bit=21 (c) saturated=19 (c') sat-source=3 (c'') low-entropy=72 (c) distinct=2
|
||||
attempt 1: 6773 candidates, rejected 4572 (67.5%): (a) stale=561 (b) no-inject=151 (a') unfresh=3774 (c) const-bit=19 (c) saturated=9 (c') sat-source=2 (c'') low-entropy=53 (c) distinct=3
|
||||
attempt 2: 4572 candidates, rejected 3137 (68.6%): (a) stale=340 (b) no-inject=103 (a') unfresh=2632 (c) const-bit=19 (c) saturated=14 (c') sat-source=2 (c'') low-entropy=27
|
||||
attempt 3: 3137 candidates, rejected 2156 (68.7%): (a) stale=265 (b) no-inject=84 (a') unfresh=1756 (c) const-bit=9 (c) saturated=9 (c') sat-source=3 (c'') low-entropy=28 (c) distinct=2
|
||||
mean accepted attempt 2.112; per-candidate rejection probability 0.6787; P(256 consecutive rejections) if independent = 7.977e-44
|
||||
closed-form distinct-item mean of accepted programs: mean 127.9503 min 123.6704
|
||||
wrote /srv/builds/_adv-adv-accept/attempts-10k-a.txt
|
||||
attempts DONE
|
||||
lease: released 32 pool cores after 941 s, exit 0
|
||||
10001
docs/analysis/cryptanalysis/logs/adv-accept/attempts-10k-a.txt
Normal file
10001
docs/analysis/cryptanalysis/logs/adv-accept/attempts-10k-a.txt
Normal file
File diff suppressed because it is too large
Load diff
95
docs/analysis/cryptanalysis/logs/adv-accept/gap-deep4.log
Normal file
95
docs/analysis/cryptanalysis/logs/adv-accept/gap-deep4.log
Normal file
|
|
@ -0,0 +1,95 @@
|
|||
lease: 0 of 4 pool cores free (88 leased); waiting
|
||||
lease: holding 8 pool cores (8,9,43,44,45,46,47,48, waited 667 s, class adv): adv-accept gap-deep4
|
||||
gap: seed 148927
|
||||
gap: program id c789fe5789ea71f0 attempt 0 class mx8-eracbf8f1c9+sh256x27 day 20730 (cache filled in 0.4 s) plant zero-dataset false
|
||||
gap: mirror validated against hash_warp on 8 units (closed form and live)
|
||||
gap: closed-form per-site distinct counts equal the library's distinct_indices_v4 on 4096 units
|
||||
gap: (c) metrics on 64 units: closed distinct_mean 128.000 saturated 0 bias_max 63; live distinct_mean 128.000 saturated 1 bias_max 69
|
||||
gap: site 0 closed ratio 0.9814 (1027095) live ratio 0.9817 (1027358) diff -0.0003; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 1 closed ratio 1.0001 (1040512) live ratio 1.0000 (1040419) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 2 closed ratio 0.9986 (1045107) live ratio 0.9988 (1045258) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 3 closed ratio 1.0000 (1046567) live ratio 1.0000 (1046556) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 4 closed ratio 1.0000 (1046483) live ratio 1.0000 (1046504) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 5 closed ratio 0.9999 (1046467) live ratio 1.0000 (1046489) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 6 closed ratio 1.0000 (1046541) live ratio 1.0000 (1046517) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 7 closed ratio 1.0000 (1044478) live ratio 1.0000 (1044528) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 8 closed ratio 1.0000 (1040340) live ratio 1.0001 (1040466) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 9 closed ratio 1.0001 (1040476) live ratio 1.0001 (1040502) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 10 closed ratio 1.0000 (1040400) live ratio 1.0001 (1040517) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 11 closed ratio 0.9999 (1046439) live ratio 1.0000 (1046523) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 12 closed ratio 1.0000 (1040381) live ratio 1.0000 (1040371) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 13 closed ratio 1.0001 (1044585) live ratio 0.9999 (1044368) diff +0.0002; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 14 closed ratio 1.0000 (1044438) live ratio 1.0000 (1044452) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 15 closed ratio 1.0000 (1044502) live ratio 1.0000 (1044510) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: min ratio closed 0.9814 live 0.9817; largest site gap -0.0003 at site 0; verdict under the 0.98 floor: closed accept live accept; 11.0 s
|
||||
gap DONE c789fe5789ea71f0
|
||||
gap: seed 150347
|
||||
gap: program id 23bc0a65fff5682b attempt 2 class mx8-era3d8e7f58+sh256x27 day 20730 (cache filled in 0.4 s) plant zero-dataset false
|
||||
gap: mirror validated against hash_warp on 8 units (closed form and live)
|
||||
gap: closed-form per-site distinct counts equal the library's distinct_indices_v4 on 4096 units
|
||||
gap: (c) metrics on 64 units: closed distinct_mean 128.000 saturated 0 bias_max 59; live distinct_mean 128.000 saturated 0 bias_max 60
|
||||
gap: site 0 closed ratio 1.0000 (1044490) live ratio 1.0000 (1044433) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 1 closed ratio 1.0001 (1040518) live ratio 1.0001 (1040500) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 2 closed ratio 1.0000 (1044499) live ratio 1.0001 (1044534) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 3 closed ratio 0.9896 (1029530) live ratio 0.9894 (1029398) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 4 closed ratio 1.0000 (1046524) live ratio 1.0000 (1046562) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 5 closed ratio 1.0000 (1044525) live ratio 0.9999 (1044387) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 6 closed ratio 1.0002 (1040541) live ratio 1.0000 (1040361) diff +0.0002; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 7 closed ratio 1.0001 (1040533) live ratio 1.0001 (1040469) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 8 closed ratio 1.0000 (1046531) live ratio 1.0000 (1046482) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 9 closed ratio 0.9998 (1044278) live ratio 0.9997 (1044187) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 10 closed ratio 0.9999 (1040319) live ratio 1.0000 (1040397) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 11 closed ratio 1.0000 (1046540) live ratio 1.0001 (1046610) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 12 closed ratio 1.0001 (1040460) live ratio 1.0001 (1040517) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 13 closed ratio 1.0000 (1046515) live ratio 0.9999 (1046450) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 14 closed ratio 1.0000 (1044439) live ratio 1.0000 (1044466) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 15 closed ratio 1.0000 (1040417) live ratio 1.0000 (1040339) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: min ratio closed 0.9896 live 0.9894; largest site gap +0.0002 at site 6; verdict under the 0.98 floor: closed accept live accept; 11.4 s
|
||||
gap DONE 23bc0a65fff5682b
|
||||
gap: seed 29307
|
||||
gap: program id 0243919aa1b29078 attempt 4 class mx8-erab0d4cf9d+sh256x27 day 20730 (cache filled in 0.4 s) plant zero-dataset false
|
||||
gap: mirror validated against hash_warp on 8 units (closed form and live)
|
||||
gap: closed-form per-site distinct counts equal the library's distinct_indices_v4 on 4096 units
|
||||
gap: (c) metrics on 64 units: closed distinct_mean 127.999 saturated 0 bias_max 60; live distinct_mean 127.999 saturated 0 bias_max 54
|
||||
gap: site 0 closed ratio 1.0001 (1040508) live ratio 1.0001 (1040447) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 1 closed ratio 0.9912 (1037356) live ratio 0.9913 (1037437) diff -0.0001; reads on indices read >= 8 times: closed 0.8769% live 0.8615%
|
||||
gap: site 2 closed ratio 1.0000 (1044500) live ratio 1.0000 (1044474) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 3 closed ratio 1.0000 (1044442) live ratio 0.9999 (1044420) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 4 closed ratio 1.0000 (1046518) live ratio 0.9999 (1046475) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 5 closed ratio 1.0001 (1044567) live ratio 1.0000 (1044505) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 6 closed ratio 1.0000 (1044492) live ratio 1.0000 (1044513) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 7 closed ratio 1.0000 (1046523) live ratio 1.0000 (1046554) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 8 closed ratio 0.9999 (1040296) live ratio 1.0000 (1040399) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 9 closed ratio 1.0000 (1040424) live ratio 1.0001 (1040458) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 10 closed ratio 1.0001 (1046625) live ratio 1.0000 (1046515) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 11 closed ratio 0.9997 (1046220) live ratio 0.9997 (1046179) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 12 closed ratio 1.0000 (1040349) live ratio 1.0001 (1040470) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 13 closed ratio 1.0000 (1044461) live ratio 1.0000 (1044529) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 14 closed ratio 1.0001 (1040500) live ratio 0.9999 (1040298) diff +0.0002; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 15 closed ratio 1.0000 (1040426) live ratio 0.9999 (1040294) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: min ratio closed 0.9912 live 0.9913; largest site gap +0.0002 at site 14; verdict under the 0.98 floor: closed accept live accept; 12.0 s
|
||||
gap DONE 0243919aa1b29078
|
||||
gap: seed 34501
|
||||
gap: program id 8b048a553cb6de4c attempt 5 class mx8-era3869ca66+sh256x27 day 20730 (cache filled in 0.4 s) plant zero-dataset false
|
||||
gap: mirror validated against hash_warp on 8 units (closed form and live)
|
||||
gap: closed-form per-site distinct counts equal the library's distinct_indices_v4 on 4096 units
|
||||
gap: (c) metrics on 64 units: closed distinct_mean 128.000 saturated 0 bias_max 67; live distinct_mean 128.000 saturated 0 bias_max 63
|
||||
gap: site 0 closed ratio 0.9929 (1033010) live ratio 0.9929 (1033029) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 1 closed ratio 1.0002 (1040630) live ratio 1.0003 (1040669) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 2 closed ratio 1.0000 (1046531) live ratio 1.0001 (1046605) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 3 closed ratio 1.0001 (1046582) live ratio 1.0000 (1046570) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 4 closed ratio 1.0000 (1044462) live ratio 1.0001 (1044603) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 5 closed ratio 1.0000 (1044498) live ratio 1.0000 (1044515) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 6 closed ratio 1.0000 (1044487) live ratio 1.0001 (1044617) diff -0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 7 closed ratio 1.0001 (1040440) live ratio 1.0000 (1040347) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 8 closed ratio 1.0000 (1046539) live ratio 1.0000 (1046519) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 9 closed ratio 1.0000 (1044471) live ratio 1.0000 (1044445) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 10 closed ratio 1.0001 (1046594) live ratio 1.0000 (1046551) diff +0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 11 closed ratio 1.0000 (1044446) live ratio 0.9999 (1044358) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 12 closed ratio 1.0000 (1044523) live ratio 1.0000 (1044451) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 13 closed ratio 1.0001 (1040468) live ratio 1.0001 (1040488) diff -0.0000; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 14 closed ratio 1.0001 (1046629) live ratio 1.0000 (1046546) diff +0.0001; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: site 15 closed ratio 0.9998 (1040207) live ratio 1.0001 (1040470) diff -0.0003; reads on indices read >= 8 times: closed 0.0000% live 0.0000%
|
||||
gap: min ratio closed 0.9929 live 0.9929; largest site gap -0.0003 at site 15; verdict under the 0.98 floor: closed accept live accept; 10.6 s
|
||||
gap DONE 8b048a553cb6de4c
|
||||
lease: released 8 pool cores after 722 s, exit 0
|
||||
|
|
@ -36,7 +36,7 @@ check is owed once I copy its program.json read-only from build-1.
|
|||
|---|---|---|---|---|---|
|
||||
| Q1 | Steering/hot set: the chain's class v4 draw over the F8 label space (shards of 100k seeds; 19:27 BST to 19:4x BST under the SIGSTOP yield, since 19:4x BST nice 10 on cores 8-95, no yield), the stand-in per-site ratio at 256 units as the proxy, the live hot set by the unmodified f8 harness (adv-live) on the lowest-ratio seeds and on a consecutive-seed census | adv-live const-item plant FLAGGED (21.8x, X_1% +6.49); clean control 0.9996x PASS | X_f >= f flags a hot set; gain = implied on-die SRAM copy size | 88,051 accepted programs drawn; eight pass the rule and flag the f8 hot-set test at 2^24 nonces (100767, 4346, 5245, 1738, 170, 106924, 107022, 122960; X_0.1% +0.102% to +0.221%, 1.54x to 2.24x), 100767 also under the class v5 dataset; 30 lowest stand-in-ratio seeds live: 18 beyond the f8 1.2x gate, 8 hot sets; 20 random: 1 beyond, 0 hot sets; every hot set is 1 MB of items holding 0.26 to 0.41% of reads: gain 1.002x; the class v5 floor at 0.995 refuses all seven read at 2^20 (0.9821 to 0.9919) | FINDING, bounded: a program-dataflow distinguisher (site 6, a near-zero source under the 1% (c') limit) that survives the class v5 dataset; not an exploitable bypass; the 256-unit selector is weak (7 of 12 false positives); shards and 23 live rows re-queued in the lease pool |
|
||||
| Q2 | Stand-in gap: a validated mirror runs the rule's own units on the closed form and on the live dataset; per-site ratio at 2^20 and the (c) metrics compared | zero-dataset plant: live 0.9733 REJECT against closed accept (fired) | any false accept that reads a hot set on the live set | 54 programs: max per-site gap 0.0004 (the 2^20 sampling noise), mean min-ratio gap 0.00002, 0 verdict disagreements, (c) metrics agree to 0.019 of 128 | BOUND |
|
||||
| Q3 | Handed to lane adv-accept-3 (unspawned); row 90 (exhaustion census: per-part rejections, cap, last resort) claimed by this lane, owner adv-accept-3 | forced-exhaust plant must hit the cap and print the last-resort program | P(exhaust) bounded; last resort characterised | attempts census over 20k seeds RUNNING (box 2); plant RUNNING | RUNNING (owner adv-accept-3, unspawned) |
|
||||
| Q3 | Row 90 (exhaustion census) claimed by this lane, owner adv-accept-3 (unspawned): every candidate of 10,000 seeds through the real rule, first failing part recorded | forced-exhaust plant hit the cap and printed the last-resort program (accepted as drawn, no lossy op) | P(exhaust) bounded; last resort characterised | 21,119 rejections: (a') 83.3%, (a) 11.7%, (b) 3.1%, (c'') 1.1%, the rest under 1%; per-candidate rejection 0.6787 flat across attempts; P(exhaust) 8e-44; 0 exhaustions; max attempt 24; second 10,000 queued | BOUND (owner adv-accept-3, unspawned) |
|
||||
| Q4 | Handed to lane adv-accept-2 | | | | HANDED OVER |
|
||||
| Q5 | Generator distinguishers over accepted programs: lossy last write, register-set collapse, the per-site ratio tail against the 0.98 floor | the five lowest-ratio seeds reproduce as live tail programs (done); F8 p15/p18/p19/p56 reproduction owed | a structural shape that lowers the live distinct-item count | first 864: lossy last write in 83%, 0 register collapse, min ratio 0.9986 at 256 units; at 2^20 the tail reaches 0.9832 (floor 0.98) | RUNNING |
|
||||
| Q6 | Fixed (c) sample grindability; live-size invariance of the (c) verdict | | BOUND unless something fires | not started | PENDING |
|
||||
|
|
@ -322,16 +322,26 @@ distinct-index ratio at the acceptance's own 2^20 sample, against the class v5 f
|
|||
| seed 19921 | fe35af21bc43a4cc | 4 | 0.9912 | refused | CLEAN live: 1.0005x (a false refusal) |
|
||||
| seed 600738 | ef6d49f3871af472 | 8 | 0.9934 | refused | CLEAN live: 1.0002x (a false refusal) |
|
||||
| seed 602822 | d8b5571c7abc80df | 3 | 0.9951 | at or above 0.995: passes | CLEAN live: 1.0003x (a true pass) |
|
||||
| seed 148927 (gap-deep4.log, box 2, 23:23 BST; the lowest 256-unit read of 143,734) | c789fe5789ea71f0 | 0 | 0.9814 | refused | CLEAN live: 1.0028x (a false refusal) |
|
||||
| seed 150347 | 23bc0a65fff5682b | 3 | 0.9896 | refused | CLEAN live: 1.0000x (a false refusal) |
|
||||
| seed 34501 | 8b048a553cb6de4c | 0 | 0.9929 | refused | CLEAN live: 0.9998x (a false refusal) |
|
||||
| seed 29307 | 0243919aa1b29078 | 1 (instr 3, src r3) | 0.9912 | refused | beyond the gate at 1.29x, no hot set by X_f; one item takes 0.027% of all reads (the third shape) |
|
||||
| Devnet 3 epoch 0, class v4 | fce15bf61030be57, 0 (day 20733) | 0 (instr 3, src r6) | 0.9992 / 0.9993 | AT OR ABOVE: a residual the floor misses | adv-cache-2: site 0 bit law P(bit 0..2 = 1) = 0.250 / 0.375 / 0.4375, chi2/dof 3.70 at 2^26 |
|
||||
| adv-cache-2 era-fixed-20 | 11f9f955b21d56c9, 1 | 11 (instr 46) | 0.9997 / 0.9997 | AT OR ABOVE: a residual | adv-cache-2: top 0.1% 1.26x its window control, one item read 512 times |
|
||||
| adv-cache-2 era-drawn-2 | 7ceb797d31eedb3e, 0 | 13 (instr 55) | 0.9992 / 0.9993 | AT OR ABOVE: a residual | adv-cache-2: chi2/dof 2.04, top 0.1% 1.27x |
|
||||
|
||||
Reading. The floor at 0.995 refuses every program whose live hot set passes the f8 test X_f >= f (all
|
||||
eight of this lane's, minimum sites 0.9821 to 0.9919). It also refuses programs with NO live concentration:
|
||||
of the five deepest 256-unit seeds, three (130488, 19921, 600738) read 0.9898 to 0.9934 at their minimum
|
||||
site and are clean on the live set (1.0002x to 1.0005x), so a site can be measurably less than uniform at
|
||||
2^20 on the stand-in without concentrating item reads live; these are the floor's false refusals, the
|
||||
population it pays as its 2.4 percent clean rejection rate, and they cost the chain nothing but a redraw.
|
||||
of the nine deepest 256-unit seeds read at 2^20, six (130488, 19921, 600738, 148927, 150347, 34501) read
|
||||
0.9814 to 0.9934 at their minimum site and are clean on the live set (0.9998x to 1.0028x of the window
|
||||
model, X_0.1% at 0.000 percent), one (602822) passes at 0.9951 and is clean, and two (122960, 29307) are
|
||||
refused and concentrate live. So the (c'') statistic, repeated word indices at one site over 2^20 random
|
||||
nonces, measures a stand-in deficiency that on the live set is usually spread thin (many indices read twice
|
||||
or three times, no item above the noise in the top 0.1 percent) and only sometimes a hot set; the six are
|
||||
the floor's false refusals, the population its 2.4 percent clean rejection rate pays, and each costs the
|
||||
chain one redraw and nothing else. The 256-unit proxy and the 2^20 read agree on these deep seeds (148927:
|
||||
0.9811 and 0.9814), so the proxy's weakness at the extreme is not sampling noise but this: a low ratio is
|
||||
not a hot set.
|
||||
The floor does not reach the three milder concentrations (top 0.1
|
||||
percent share 1.26x to 1.27x, chi2/dof 2 to 3.7) whose minimum site reads 0.9992 to 0.9997, inside the clean
|
||||
population's spread at 2^20 (clean median 0.9999, clean first percentile about 0.9993 over 29,032 programs).
|
||||
|
|
@ -341,6 +351,35 @@ they are distinguishers at the chi-square level, not hot sets, and no floor on t
|
|||
keeps the clean rejection rate near 2.4 percent reaches them. The 1738 hot set's ratio (gap-hot3 on build-1)
|
||||
lands when its lease gets cores.
|
||||
|
||||
### Row 90 result (owner adv-accept-3, unspawned; claimed by this lane): the attempt loop over 10,000 seeds (box 2, 23:22 BST)
|
||||
|
||||
`adv-accept attempts --seed-start 2 --seeds 10000 --threads 32 --out attempts-10k-a.txt` through
|
||||
`lease pool 32 --min 16` (log logs/adv-accept/attempts-10k-a.log, rows attempts-10k-a.txt): every candidate
|
||||
of every seed walked through the real rule, the first failing part recorded; 940 s on 32 cores. The second
|
||||
half (seeds 10002 to 20001) waits on build-1.
|
||||
|
||||
| Measure | Value |
|
||||
|---|---|
|
||||
| Seeds, accepted, exhausted | 10,000, 10,000, 0 (cap 256) |
|
||||
| Rejected candidates | 21,119; per-candidate rejection probability 0.6787 |
|
||||
| Rejection by part, all candidates | (a') unfresh 83.3%, (a) stale 11.7%, (b) no injecting write 3.1%, (c'') low-entropy site 1.1%, (c) constant bit 0.4%, (c) saturated 0.3%, (c') saturated source 0.1%, (c) distinct addresses 0.04%, (c) lane-constant 0, (c) bias 0, (c'') repeated 0 (not wired) |
|
||||
| Attempt 0 | 10,000 candidates, 6,773 rejected (67.7%) |
|
||||
| Attempts 1, 2, 3 | 67.5%, 68.6%, 68.7% rejected: the per-attempt rate does not drift |
|
||||
| Accepted-attempt histogram 0..11 | 3,227, 2,201, 1,435, 981, 732, 449, 308, 211, 137, 118, 64, 41; max 24 (one seed); mean 2.112 |
|
||||
| P(256 consecutive rejections) if attempts are independent | 0.6787^256 = 8.0e-44 |
|
||||
| Closed-form distinct-item mean of the accepted programs | 127.95 of 128; minimum 123.67 (the floor is 120) |
|
||||
|
||||
Reading. Rule (a') does the work: five of six rejections are the dataflow freshness check at the draw's
|
||||
boundary, (a) and (b) one in seven, and the dynamic parts under 2 percent together, with (c'') the largest
|
||||
of them at 1.1 percent (238 candidates, the per-site floor at 0.98). The per-attempt rejection rate is flat
|
||||
at 68 percent across attempts 0 to 3, so attempts are independent draws and the cap of 256 is unreachable
|
||||
in practice (8e-44 per seed; the devnet at one epoch an hour would need 10^39 years). The last resort is
|
||||
therefore a path no chain reaches by chance; it is reachable only by a rule that rejects everything (the
|
||||
plant above), and what it hands over is a program with no lossy op that the rule accepts as drawn. For an
|
||||
attacker who steers the seed: the attempt index adds no freedom beyond the seed (each attempt is a fresh
|
||||
seed-word set), and the accepted attempt's distribution is the same for every seed, so nothing in the
|
||||
attempt scheme is a lever. Owner adv-accept-3 (unspawned); rows 91 and 92 are theirs.
|
||||
|
||||
### Q2, the stand-in gap: first bound (box 2, 19:46 BST, `adv-accept gap`)
|
||||
|
||||
Tool: a mirror interpreter with per-site index capture, run on the rule's own base nonces (seed-keyed
|
||||
|
|
|
|||
Loading…
Reference in a new issue