From b141231d91d4d893109b0c4e86377ea5a0e54fef Mon Sep 17 00:00:00 2001 From: igneum-labs <337424239+igneum-labs@users.noreply.github.com> Date: Thu, 8 Oct 2026 00:03:53 +0000 Subject: [PATCH] adv-cache-2: devnet 2^25 under v2, ledger (internal adversarial pass, not an independent review) Co-Authored-By: Claude Fable 5.1 --- .../logs/adv-cache-2/warps-devnet-2e25-v2.log | 65 +++++++++++++++++++ .../cryptanalysis/report-chained-cache-2.md | 8 ++- 2 files changed, 72 insertions(+), 1 deletion(-) create mode 100644 docs/analysis/cryptanalysis/logs/adv-cache-2/warps-devnet-2e25-v2.log diff --git a/docs/analysis/cryptanalysis/logs/adv-cache-2/warps-devnet-2e25-v2.log b/docs/analysis/cryptanalysis/logs/adv-cache-2/warps-devnet-2e25-v2.log new file mode 100644 index 000000000..55e139202 --- /dev/null +++ b/docs/analysis/cryptanalysis/logs/adv-cache-2/warps-devnet-2e25-v2.log @@ -0,0 +1,65 @@ +lease: holding 32 pool cores (10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41, waited 0 s, class adv): adv-cache-2 warps-devnet-2e25-v2 +[2026-10-07T22:22:24Z] adv-cache-2 0.1.0 (igneum-pow generator v4); args ["warps", "--programs", "devnet", "--day", "20730", "--nonces", "33554432", "--validate", "sample", "--threads", "32", "--out", "/srv/builds/_adv-adv-cache-2"] +[2026-10-07T22:22:24Z] warps: programs ["devnet"] day 20730 nonces 33554432 threads 32 plant none validate sample check_every 997 fingerprint false +[2026-10-07T22:22:24Z] day 20730: cache filled in 0.41 s, fnv 448274a57f508cbc +[2026-10-07T22:22:26Z] table: 16777216 items derived with their 8 lines in 1.6 s; library comparison sample (0 mismatches) +[2026-10-07T22:22:28Z] static devnet-epoch0: s0:instr1:r7:base-writer rotl@0:shadow-writes add=5,mad=5,mul=4,mulhi=1,rotr=4,shfl=2,sub=3,xor=3:last-shadow-writer add | s1:instr4:r4:base-writer load@1:shadow-writes add=7,mad=1,mul=7,mulhi=2,or=5,rotr=1,shfl=5,sub=2,xor=5:last-shadow-writer or | s2:instr6:r0:base-writer load@4:shadow-writes add=8,mad=9,mul=4,mulhi=2,rotl=6,rotr=2,shfl=2,sub=3,xor=8:last-shadow-writer add | s3:instr10:r7:base-writer mad@2:shadow-writes add=5,mad=5,mul=4,mulhi=1,rotr=4,shfl=2,sub=3,xor=3:last-shadow-writer add | s4:instr12:r2:base-writer rotl@11:shadow-writes add=4,mad=3,mul=3,mulhi=3,or=1,rotl=4,rotr=3,shfl=2,sub=4,xor=5:last-shadow-writer add | s5:instr20:r1:base-writer shfl@18:shadow-writes add=5,mad=3,mulhi=1,or=3,rotl=1,rotr=3,shfl=5,sub=7,xor=3:last-shadow-writer rotr | s6:instr27:r7:base-writer mad@14:shadow-writes add=5,mad=5,mul=4,mulhi=1,rotr=4,shfl=2,sub=3,xor=3:last-shadow-writer add | s7:instr30:r2:base-writer sub@19:shadow-writes add=4,mad=3,mul=3,mulhi=3,or=1,rotl=4,rotr=3,shfl=2,sub=4,xor=5:last-shadow-writer add | s8:instr35:r4:base-writer xor@33:shadow-writes add=7,mad=1,mul=7,mulhi=2,or=5,rotr=1,shfl=5,sub=2,xor=5:last-shadow-writer or | s9:instr40:r7:base-writer mad@38:shadow-writes add=5,mad=5,mul=4,mulhi=1,rotr=4,shfl=2,sub=3,xor=3:last-shadow-writer add | s10:instr41:r0:base-writer mad@21:shadow-writes add=8,mad=9,mul=4,mulhi=2,rotl=6,rotr=2,shfl=2,sub=3,xor=8:last-shadow-writer add | s11:instr43:r3:base-writer sub@22:shadow-writes add=6,mad=2,mul=2,mulhi=2,or=3,rotl=1,rotr=2,shfl=2,sub=2,xor=3:last-shadow-writer shfl | s12:instr45:r1:base-writer load@43:shadow-writes add=5,mad=3,mulhi=1,or=3,rotl=1,rotr=3,shfl=5,sub=7,xor=3:last-shadow-writer rotr | s13:instr52:r1:base-writer rotr@46:shadow-writes add=5,mad=3,mulhi=1,or=3,rotl=1,rotr=3,shfl=5,sub=7,xor=3:last-shadow-writer rotr | s14:instr53:r3:base-writer load@52:shadow-writes add=6,mad=2,mul=2,mulhi=2,or=3,rotl=1,rotr=2,shfl=2,sub=2,xor=3:last-shadow-writer shfl | s15:instr54:r5:base-writer load@53:shadow-writes add=5,mad=6,mul=3,mulhi=2,or=3,rotl=6,rotr=3,shfl=3,xor=7:last-shadow-writer xor +[2026-10-07T22:22:28Z] program devnet-epoch0: epoch seed edc4fa844da9dc98d37e965176f6558a31560e40502ab3ae5491b21aaaabfb07 era seed edc4fa844da9dc98d37e965176f6558a31560e40502ab3ae5491b21aaaabfb07 id a785001687d8688a attempt 1 class mx8-erad810f22d+sh256x27 op mix load=16 mad=8 mul=6 xor=6 rotr=5 shfl=5 sub=5 add=4 mulhi=4 rotl=4 or=1; era stride mul 0x9ad30d99 rot 29 interleave [0, 2, 12, 13]; sites instr:win:off 1:2:3 4:2:1 6:2:1 10:1:1 12:0:0 20:0:0 27:0:0 30:0:0 35:2:3 40:0:0 41:2:1 43:0:0 45:2:1 52:1:1 53:0:0 54:2:1 +[2026-10-07T22:22:28Z] window layer devnet-epoch0: site windows (first, items) (0xc00000,2^22) (0x400000,2^22) (0x400000,2^22) (0x800000,2^23) (0x0,2^24) (0x0,2^24) (0x0,2^24) (0x0,2^24) (0xc00000,2^22) (0x0,2^24) (0x400000,2^22) (0x0,2^24) (0x400000,2^22) (0x800000,2^23) (0x0,2^24) (0x400000,2^22); expected share per quarter 0.1094 0.4219 0.1719 0.2969; top quarter 0.4219 (uniform 0.25), top aligned half 0.5312 (uniform 0.5) +[2026-10-07T22:23:02Z] devnet-epoch0: 1048576 warps (33554432 nonces) interpreted in 33.3 s (1.016 ms per warp per thread); Epoch::hash_warp agreement on 1115 warps: 0 mismatches +[2026-10-07T22:23:02Z] devnet-epoch0: warp 0 lanes agree with the pack's vectors.json +[2026-10-07T22:23:02Z] devnet-epoch0 item histogram (flat): bins 16777216 reads 4294967296 mean 256.000 sigma 16.000 max 539 (bin 6539211) z_max +17.69 min 64 z_min -12.00 chi2/dof 60.00317 chi2_z +170891.42 top0.1% 0.19332% top0.5% 0.94360% top1% 1.86482% +[2026-10-07T22:23:02Z] devnet-epoch0: measured share per quarter 0.10937 0.42187 0.17188 0.29689; top quarter 0.42187 (window model 0.42188, uniform 0.25); top aligned half 0.53124 (model 0.53125, uniform 0.5) +[2026-10-07T22:23:06Z] devnet-epoch0 WINDOWED CONTROL item histogram: bins 16777216 reads 4294967296 mean 256.000 sigma 16.000 max 539 (bin 5577539) z_max +17.69 min 64 z_min -12.00 chi2/dof 59.99640 chi2_z +170871.83 top0.1% 0.19326% top0.5% 0.94356% top1% 1.86474% +[2026-10-07T22:23:09Z] devnet-epoch0 FLAT CONTROL item histogram: bins 16777216 reads 4294967296 mean 256.000 sigma 16.000 max 344 (bin 11889943) z_max +5.50 min 176 z_min -5.00 chi2/dof 1.00006 chi2_z +0.16 top0.1% 0.12169% top0.5% 0.59275% top1% 1.17056% +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (windowed control): f 0.1% S_f 0.19332% E_f(control) 0.19326% X_f +0.00005% X_f/f +0.0005 ratio S/E 1.0003x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (windowed control): f 0.5% S_f 0.94360% E_f(control) 0.94356% X_f +0.00004% X_f/f +0.0001 ratio S/E 1.0000x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (windowed control): f 1.0% S_f 1.86482% E_f(control) 1.86474% X_f +0.00008% X_f/f +0.0001 ratio S/E 1.0000x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (windowed control): verdict clear +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (flat control): f 0.1% S_f 0.19332% E_f(control) 0.12169% X_f +0.07162% X_f/f +0.7162 ratio S/E 1.5885x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (flat control): f 0.5% S_f 0.94360% E_f(control) 0.59275% X_f +0.35086% X_f/f +0.7017 ratio S/E 1.5919x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (flat control): f 1.0% S_f 1.86482% E_f(control) 1.17056% X_f +0.69426% X_f/f +0.6943 ratio S/E 1.5931x -> no hot set +[2026-10-07T22:23:09Z] hot-set devnet-epoch0 items (flat control): verdict clear +[2026-10-07T22:23:09Z] 6-sigma devnet-epoch0 items buckets64 against the window density: largest +4.55 sigma smallest -4.02; WINDOWED CONTROL largest +4.76 smallest -4.69 -> within 6 sigma +[2026-10-07T22:23:10Z] site 0 (instr 1, win 2, off 3, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 110 (bin 1720924) z_max +5.75 min 28 z_min -4.50 chi2/dof 1.01575 chi2_z +22.81 top0.1% 0.14522% top0.5% 0.69216% top1% 1.35278%; control top0.1% 0.14473% top1% 1.34882%; ratio top0.1% 1.0033x top1% 1.0029x; z_max +5.75 +[2026-10-07T22:23:10Z] site 1 (instr 4, win 2, off 1, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 106 (bin 3230518) z_max +5.25 min 28 z_min -4.50 chi2/dof 1.00310 chi2_z +4.49 top0.1% 0.14459% top0.5% 0.69057% top1% 1.34945%; control top0.1% 0.14461% top1% 1.34840%; ratio top0.1% 0.9999x top1% 1.0008x; z_max +5.25 +[2026-10-07T22:23:10Z] site 2 (instr 6, win 2, off 1, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 113 (bin 465483) z_max +6.12 min 28 z_min -4.50 chi2/dof 0.99905 chi2_z -1.37 top0.1% 0.14463% top0.5% 0.69018% top1% 1.34844%; control top0.1% 0.14454% top1% 1.34804%; ratio top0.1% 1.0007x top1% 1.0003x; z_max +6.12 +[2026-10-07T22:23:11Z] site 3 (instr 10, win 1, off 1, window 2^23 items, reads outside the window 0): hist: bins 8388608 reads 268435456 mean 32.000 sigma 5.657 max 65 (bin 1490381) z_max +5.83 min 7 z_min -4.42 chi2/dof 0.99950 chi2_z -1.03 top0.1% 0.16460% top0.5% 0.77436% top1% 1.50124%; control top0.1% 0.16465% top1% 1.50148%; ratio top0.1% 0.9997x top1% 0.9998x; z_max +5.83 +[2026-10-07T22:23:11Z] site 4 (instr 12, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 40 (bin 12663392) z_max +6.00 min 0 z_min -4.00 chi2/dof 0.99864 chi2_z -3.93 top0.1% 0.19403% top0.5% 0.89625% top1% 1.72392%; control top0.1% 0.19429% top1% 1.72504%; ratio top0.1% 0.9987x top1% 0.9993x; z_max +6.00 +[2026-10-07T22:23:12Z] site 5 (instr 20, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 41 (bin 7223125) z_max +6.25 min 0 z_min -4.00 chi2/dof 0.99976 chi2_z -0.69 top0.1% 0.19413% top0.5% 0.89644% top1% 1.72416%; control top0.1% 0.19415% top1% 1.72467%; ratio top0.1% 0.9999x top1% 0.9997x; z_max +6.25 +[2026-10-07T22:23:13Z] site 6 (instr 27, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 41 (bin 11613008) z_max +6.25 min 0 z_min -4.00 chi2/dof 1.01903 chi2_z +55.13 top0.1% 0.19543% top0.5% 0.90290% top1% 1.73439%; control top0.1% 0.19433% top1% 1.72476%; ratio top0.1% 1.0057x top1% 1.0056x; z_max +6.25 +[2026-10-07T22:23:14Z] site 7 (instr 30, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 45 (bin 3079588) z_max +7.25 min 0 z_min -4.00 chi2/dof 0.99991 chi2_z -0.27 top0.1% 0.19429% top0.5% 0.89696% top1% 1.72480%; control top0.1% 0.19428% top1% 1.72450%; ratio top0.1% 1.0001x top1% 1.0002x; z_max +7.25 +[2026-10-07T22:23:14Z] site 8 (instr 35, win 2, off 3, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 112 (bin 230572) z_max +6.00 min 26 z_min -4.75 chi2/dof 1.00061 chi2_z +0.89 top0.1% 0.14464% top0.5% 0.69040% top1% 1.34907%; control top0.1% 0.14454% top1% 1.34826%; ratio top0.1% 1.0007x top1% 1.0006x; z_max +6.00 +[2026-10-07T22:23:15Z] site 9 (instr 40, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 42 (bin 7701642) z_max +6.50 min 0 z_min -4.00 chi2/dof 0.99972 chi2_z -0.80 top0.1% 0.19423% top0.5% 0.89673% top1% 1.72450%; control top0.1% 0.19437% top1% 1.72419%; ratio top0.1% 0.9993x top1% 1.0002x; z_max +6.50 +[2026-10-07T22:23:15Z] site 10 (instr 41, win 2, off 1, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 109 (bin 2908981) z_max +5.62 min 28 z_min -4.50 chi2/dof 1.00004 chi2_z +0.06 top0.1% 0.14470% top0.5% 0.69060% top1% 1.34920%; control top0.1% 0.14465% top1% 1.34938%; ratio top0.1% 1.0004x top1% 0.9999x; z_max +5.62 +[2026-10-07T22:23:16Z] site 11 (instr 43, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 42 (bin 8532382) z_max +6.50 min 0 z_min -4.00 chi2/dof 0.99990 chi2_z -0.30 top0.1% 0.19425% top0.5% 0.89703% top1% 1.72479%; control top0.1% 0.19426% top1% 1.72482%; ratio top0.1% 1.0000x top1% 1.0000x; z_max +6.50 +[2026-10-07T22:23:16Z] site 12 (instr 45, win 2, off 1, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 109 (bin 1447810) z_max +5.62 min 29 z_min -4.38 chi2/dof 1.00018 chi2_z +0.26 top0.1% 0.14466% top0.5% 0.69029% top1% 1.34876%; control top0.1% 0.14448% top1% 1.34820%; ratio top0.1% 1.0013x top1% 1.0004x; z_max +5.62 +[2026-10-07T22:23:17Z] site 13 (instr 52, win 1, off 1, window 2^23 items, reads outside the window 0): hist: bins 8388608 reads 268435456 mean 32.000 sigma 5.657 max 67 (bin 1574762) z_max +6.19 min 7 z_min -4.42 chi2/dof 0.99990 chi2_z -0.21 top0.1% 0.16464% top0.5% 0.77437% top1% 1.50133%; control top0.1% 0.16455% top1% 1.50117%; ratio top0.1% 1.0005x top1% 1.0001x; z_max +6.19 +[2026-10-07T22:23:18Z] site 14 (instr 53, win 0, off 0, window 2^24 items, reads outside the window 0): hist: bins 16777216 reads 268435456 mean 16.000 sigma 4.000 max 45 (bin 10096881) z_max +7.25 min 0 z_min -4.00 chi2/dof 1.00045 chi2_z +1.29 top0.1% 0.19432% top0.5% 0.89718% top1% 1.72516%; control top0.1% 0.19434% top1% 1.72491%; ratio top0.1% 0.9999x top1% 1.0001x; z_max +7.25 +[2026-10-07T22:23:18Z] site 15 (instr 54, win 2, off 1, window 2^22 items, reads outside the window 0): hist: bins 4194304 reads 268435456 mean 64.000 sigma 8.000 max 113 (bin 1198078) z_max +6.12 min 28 z_min -4.50 chi2/dof 1.00092 chi2_z +1.33 top0.1% 0.14466% top0.5% 0.69029% top1% 1.34873%; control top0.1% 0.14468% top1% 1.34871%; ratio top0.1% 0.9998x top1% 1.0000x; z_max +6.12 +[2026-10-07T22:23:19Z] devnet-epoch0 LINE histogram (8 lines per item read): bins 4194304 reads 34359738368 mean 8192.000 sigma 90.510 max 17509 (bin 1956659) z_max +102.94 min 1855 z_min -70.01 chi2/dof 315.64444 chi2_z +455653.77 top0.1% 0.17338% top0.5% 0.81017% top1% 1.56684% +[2026-10-07T22:23:20Z] devnet-epoch0 LINE histogram under the WINDOWED CONTROL items: bins 4194304 reads 34359738368 mean 8192.000 sigma 90.510 max 17631 (bin 3638725) z_max +104.29 min 1830 z_min -70.29 chi2/dof 315.64928 chi2_z +455660.78 top0.1% 0.17338% top0.5% 0.81015% top1% 1.56685% +[2026-10-07T22:23:20Z] hot-set devnet-epoch0 lines (windowed control): f 0.1% S_f 0.17338% E_f(control) 0.17338% X_f -0.00000% X_f/f -0.0000 ratio S/E 1.0000x -> no hot set +[2026-10-07T22:23:20Z] hot-set devnet-epoch0 lines (windowed control): f 0.5% S_f 0.81017% E_f(control) 0.81015% X_f +0.00002% X_f/f +0.0000 ratio S/E 1.0000x -> no hot set +[2026-10-07T22:23:20Z] hot-set devnet-epoch0 lines (windowed control): f 1.0% S_f 1.56684% E_f(control) 1.56685% X_f -0.00000% X_f/f -0.0000 ratio S/E 1.0000x -> no hot set +[2026-10-07T22:23:20Z] hot-set devnet-epoch0 lines (windowed control): verdict clear +[2026-10-07T22:23:20Z] devnet-epoch0 SEGMENT histogram (lines / 64): bins 65536 reads 34359738368 mean 524288.000 sigma 724.077 max 589682 (bin 21859) z_max +90.31 min 471091 z_min -73.47 chi2/dof 316.39319 chi2_z +57091.83 top0.1% 0.10911% top0.5% 0.53623% top1% 1.06500% +[2026-10-07T22:23:20Z] devnet-epoch0 SEGMENT histogram under the WINDOWED CONTROL items: bins 65536 reads 34359738368 mean 524288.000 sigma 724.077 max 588795 (bin 21859) z_max +89.09 min 471411 z_min -73.03 chi2/dof 316.48196 chi2_z +57107.90 top0.1% 0.10909% top0.5% 0.53621% top1% 1.06505% +[2026-10-07T22:23:20Z] lines-vs-control devnet-epoch0: chi2/dof lines 315.6444 vs control 315.6493; z_max +102.94 vs +104.29; segments chi2/dof 316.3932 vs 316.4820; z_max +90.31 vs +89.09 -> matches the control +[2026-10-07T22:23:23Z] devnet-epoch0 REAL LINE STORE: f=1/2 stride: hit 0.50006 evals/item 3.9995 ops/item 2432 (+0.260x of 9360); f=1/2 hottest: hit 0.57824 evals/item 6.6365 ops/item 4035 (+0.431x of 9360); f=1/4 stride: hit 0.25005 evals/item 11.9994 ops/item 7296 (+0.779x of 9360); f=1/4 hottest: hit 0.31389 evals/item 20.9025 ops/item 12709 (+1.358x of 9360); f=1/8 stride: hit 0.12501 evals/item 27.9988 ops/item 17023 (+1.819x of 9360); f=1/8 hottest: hit 0.16709 evals/item 47.5117 ops/item 28887 (+3.086x of 9360); f=1/16 stride: hit 0.06251 evals/item 60.0003 ops/item 36480 (+3.897x of 9360); f=1/16 hottest: hit 0.08798 evals/item 89.7704 ops/item 54580 (+5.831x of 9360); f=1/32 stride: hit 0.03123 evals/item 124.0092 ops/item 75398 (+8.055x of 9360); f=1/32 hottest: hit 0.04599 evals/item 141.3220 ops/item 85924 (+9.180x of 9360); f=1/64 stride: hit 0.01561 evals/item 252.0162 ops/item 153226 (+16.370x of 9360); f=1/64 hottest: hit 0.02392 evals/item 187.3374 ops/item 113901 (+12.169x of 9360) +[2026-10-07T22:23:26Z] devnet-epoch0 WINDOWED CONTROL LINE STORE: f=1/2 stride: hit 0.50005 evals/item 3.9996 ops/item 2432 (+0.260x of 9360); f=1/2 hottest: hit 0.57824 evals/item 6.6379 ops/item 4036 (+0.431x of 9360); f=1/4 stride: hit 0.25005 evals/item 11.9995 ops/item 7296 (+0.779x of 9360); f=1/4 hottest: hit 0.31389 evals/item 20.8874 ops/item 12700 (+1.357x of 9360); f=1/8 stride: hit 0.12501 evals/item 27.9988 ops/item 17023 (+1.819x of 9360); f=1/8 hottest: hit 0.16709 evals/item 47.4760 ops/item 28865 (+3.084x of 9360); f=1/16 stride: hit 0.06251 evals/item 60.0001 ops/item 36480 (+3.897x of 9360); f=1/16 hottest: hit 0.08798 evals/item 89.7325 ops/item 54557 (+5.829x of 9360); f=1/32 stride: hit 0.03124 evals/item 124.0088 ops/item 75397 (+8.055x of 9360); f=1/32 hottest: hit 0.04599 evals/item 141.3337 ops/item 85931 (+9.181x of 9360); f=1/64 stride: hit 0.01561 evals/item 252.0167 ops/item 153226 (+16.370x of 9360); f=1/64 hottest: hit 0.02392 evals/item 187.3327 ops/item 113898 (+12.169x of 9360) +[2026-10-07T22:23:27Z] devnet-epoch0 ITEM STORE (hottest items, read-weighted): f=0.001: S_f 0.00193 (control 0.00193, uniform 0.00100); ops/hash 1196276 vs model 1197394 (0.9991x); f=0.01: S_f 0.01865 (control 0.01865, uniform 0.01000); ops/hash 1176250 vs model 1186611 (0.9913x); f=0.1: S_f 0.17661 (control 0.17660, uniform 0.10000); ops/hash 987001 vs model 1078784 (0.9149x); f=0.25: S_f 0.42187 (control 0.42187, uniform 0.25000); ops/hash 693155 vs model 899072 (0.7710x); f=0.5: S_f 0.71876 (control 0.71874, uniform 0.50000); ops/hash 337465 vs model 599552 (0.5629x); f=0.75: S_f 0.89066 (control 0.89065, uniform 0.75000); ops/hash 131511 vs model 300032 (0.4383x) +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 0: 16777216 nonces; item hist chi2/dof 1.0010 z_max +10.00 top0.1% 0.56950% top1% 4.43785%; source hi16: max bucket share 0.00196% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2502 0.3751 0.4375 0.4688 0.4843 0.4923 0.4959 0.4980 +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 1: 16777216 nonces; item hist chi2/dof 1.0010 z_max +9.00 top0.1% 0.56934% top1% 4.44008%; source hi16: max bucket share 0.00194% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2499 0.3751 0.4377 0.4689 0.4845 0.4922 0.4962 0.4980 +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 2: 16777216 nonces; item hist chi2/dof 1.0014 z_max +9.00 top0.1% 0.56974% top1% 4.43965%; source hi16: max bucket share 0.00195% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2501 0.3751 0.4376 0.4688 0.4846 0.4922 0.4962 0.4981 +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 3: 16777216 nonces; item hist chi2/dof 1.0010 z_max +10.00 top0.1% 0.56888% top1% 4.43539%; source hi16: max bucket share 0.00193% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2501 0.3750 0.4376 0.4688 0.4844 0.4922 0.4960 0.4980 +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 4: 16777216 nonces; item hist chi2/dof 1.0016 z_max +9.00 top0.1% 0.57043% top1% 4.44074%; source hi16: max bucket share 0.00198% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2499 0.3752 0.4374 0.4687 0.4845 0.4921 0.4960 0.4980 +[2026-10-07T22:23:46Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 5: 16777216 nonces; item hist chi2/dof 1.0011 z_max +10.00 top0.1% 0.57033% top1% 4.43848%; source hi16: max bucket share 0.00193% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2500 0.3749 0.4374 0.4687 0.4843 0.4921 0.4961 0.4980 +[2026-10-07T22:23:47Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 6: 16777216 nonces; item hist chi2/dof 1.0009 z_max +9.00 top0.1% 0.56969% top1% 4.43609%; source hi16: max bucket share 0.00197% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2501 0.3748 0.4375 0.4689 0.4843 0.4920 0.4961 0.4981 +[2026-10-07T22:23:47Z] diag devnet-epoch0 worst site 6 (instr 27) iteration 7: 16777216 nonces; item hist chi2/dof 1.0017 z_max +9.00 top0.1% 0.57042% top1% 4.43918%; source hi16: max bucket share 0.00194% (uniform 0.00153%) entropy 15.997 of 16 bits; saturated source 0.000000%; P(source bit b = 1) for b = 0..7: 0.2499 0.3749 0.4375 0.4687 0.4843 0.4920 0.4960 0.4980 +[2026-10-07T22:23:47Z] hottest item devnet-epoch0 (0x63c7cb): 539 reads (0.00001% of all; uniform mean 256.000) +[2026-10-07T22:23:47Z] program devnet-epoch0 DONE: 33554432 nonces; 6-sigma (windowed) clear; hot-set items clear; hot-set lines clear; verdict PASS (no test fired) +[2026-10-07T22:23:47Z] SEED devnet-epoch0 id a785001687d8688a attempt 1 S0.1 0.19332% window 0.19326% flat 0.12169% ratio 1.0003x; S1 1.86482% window 1.86474% ratio 1.0000x; top quarter 0.42187 (model 0.42188) top half 0.53124 (model 0.53125); lines S0.1 0.17338% E 0.17338% S1 1.56684% E 1.56685%; worst site 6 ratio 1.0057x; hot 0 +[2026-10-07T22:23:47Z] CENSUS PASS: 1 programs at 33554432 nonces; hot set in 0; top 0.1% beyond 1.2x of the window-model control in 0; windowed buckets64 beyond 6 sigma in 0; top-quarter share min 0.4219 median 0.4219 mean 0.4219 max 0.4219 (uniform 0.25) +Terminated +lease: released 32 pool cores after 84 s, exit 0 diff --git a/docs/analysis/cryptanalysis/report-chained-cache-2.md b/docs/analysis/cryptanalysis/report-chained-cache-2.md index b3181d641..942148f3d 100644 --- a/docs/analysis/cryptanalysis/report-chained-cache-2.md +++ b/docs/analysis/cryptanalysis/report-chained-cache-2.md @@ -242,6 +242,10 @@ Log: `logs/adv-cache-2/warps-devnet-2e26.log`. Validation: 2,167 warps vs `Epoch Reading: the first real program is the window model to 1.0002x; two of its sites carry a 3 to 4 percent excess variance at the item level (chi2/dof 1.03 to 1.04, far under Devnet 3's 1.67), worth under 0.01 percent of a hash's reads to a store. +### 2.4 Shared devnet at 2^25 nonces under the v2 pricing binary (item 430, 22:22Z to 22:23Z, 33 s of interpretation on 32 cores) + +Log: `logs/adv-cache-2/warps-devnet-2e25-v2.log`. 1,115 warps vs `Epoch::hash_warp`, 0 mismatches. Items top 0.1 percent 1.0003x of the windowed control, top 1 percent 1.0000x; lines 1.0000x; worst site 6 at 1.0057x; verdict PASS. Stores at the measured distribution: stride f = 1/2 hit 0.50006, 3.9995 evaluations per item (2,432 ops, +0.26x); hottest-lines f = 1/2 hit 0.57824, 6.6365 evaluations (4,035 ops, +0.43x); the windowed control gives the same to the fourth digit. Items store: f = 0.001 serves 0.00193 of reads (control 0.00193), f = 0.01 0.01865 (0.01865), f = 0.1 0.17661 (0.17660), f = 0.25 0.4218: the whole excess over f is the window layer, none of it the hash. + ## 3. Q3: steering ### 3.1 By the item index t (box 1, items 11 and 12, 3 s and 5 s) @@ -362,9 +366,10 @@ The first load of a hash (site 0 in iteration 0) depends only on the nonce, the | Q2b era-drawn 17 to 20 (shard s4, 176 s wall) | 2 | 176 s | 32 | 0.02 | | Q2b era-drawn 21 to 24 (shard s5, 172 s) | 2 | 172 s | 32 | 0.02 | | Q2b era-drawn 25 to 32 (shards s6, s7, about 175 s each) | 2 | 350 s | 32 | 0.03 | +| Q2 devnet 2^25 under v2 (item 430) | 2 | 84 s | 32 | 0.01 | | Pod-hours | none (no GPU, no rented pod) | | | 0 | -Running total at 22:25Z: about 2.2 box-hours by wall x threads / 96 (the boxes ran at load 400 to 600 for the first two hours, so the CPU actually consumed is under that). The 8-hour reading is not near. The 8-hour reading is not near. +Running total at 00:05Z (8 October): about 2.2 box-hours of run by wall x threads / 96; the lease waits (box 1's line shard s2c has waited since 22:09Z with 0 free adv-class cores) are not counted by wall x threads / 96 (the boxes ran at load 400 to 600 for the first two hours, so the CPU actually consumed is under that). The 8-hour reading is not near. The 8-hour reading is not near. ## 6. The bound reached, honestly @@ -383,6 +388,7 @@ Not a reason to wait: the 2^35-read census (item 16) and the 16,384-day scan (it | 19:49Z | Coordinator's kills (recorded as theirs): on box 2 the resumed census `warps-era-fixed-12-32-2e23` (lease 40 cores, taken 19:41:36Z) ended by SIGTERM after programs 12 to 23 (lost: programs 24 to 32, about 7 min; program 24's rows partial), then the drawn-era census `warps-era-drawn-32-2e23` that the still-alive box 2 drain had started on 40 cores, after programs 1 and 2 (lost: 3 to 32). The order: box 2 closed to adv-* while the class v5 census ran (19:4x to 20:12Z). My 18:2x line "both drains stopped" was wrong for box 2: the lease-aware drain restarted at 19:23:53Z was alive and started the next item on the kill. Box 1 untouched (the 2^35 census ran to its end, 19:35:36Z to 20:05:06Z) | | 20:33Z | My own wrong kills on box 1: shard `lines-2e30-s1` killed on purpose (its 64 days duplicated the 256-day census) and, by a `pkill -x` that followed, the shard `s2b` the drain had just started (1 s lost, re-added as `s2c`). Line shards s3b, s4, s2c now cover days 21002 to 21193, fresh | | 21:21Z to 21:27Z | Pre-emption rule (a release or v5 waiter pre-empts adv holders at any size after 120 s): the drain re-queues a shard that exits 143, up to 6 times, after the yield test; redeployed on box 1 at 21:23Z (waiting on the running shard s4) and box 2 at 21:27Z (waiting on the running drawn-era shard); each shard's harness logs its rows as it goes | +| 22:23Z | Box 2 queue drained (the drain exited after 6 items; the three adv-cache-3 queue files there are not lease-pool items and were skipped, not run). Box 1: the 8 drawn-era windows shards wait behind the line shard s2c, itself waiting for adv-class cores since 22:09Z; at 00:0xZ the windows shards were copied to box 2 and a drain started there | | 20:40Z | My own wrong kill: on box 2 I ended lease pid 2674115 (24 cores, label adv-cache-2 warps-devnet-2e26-v2, held from 20:36:59Z) as an orphan; it was a sibling's script running my old queue item 43 under the sibling-queue rule (claimed by my first drain, superseded by my item 430). About 3 minutes of a sibling's cores lost; reported to the coordinator at 20:42Z | | 20:2x to 20:3xZ | Box 2 reopened at 20:12Z. Remaining work re-sharded (lines 2^30 x 4 shards, windows 512 x 8, warps 4 to 5 programs per shard, devnet at 2^25) with every lease at `lease pool 32 --min 16` (holders at 32 or under are never pre-empted under the ranked lease of 19:40Z), the drain holding before each lease while a class-v5 waiter (by owner, never an adv-* lane) is in `lease status`; a first version of that test matched a v5 HOLDER and held box 1 idle from 19:20Z to 20:2xZ (fixed: a waiter line carries "(min N, K free)"); duplicate drains and an orphan 48-core waiter on box 2 killed at 20:2xZ, one drain per box since | | 19:23:49Z (box 1), 19:23:53Z (box 2) | Re-queued through `lease pool` (present on both boxes from 19:22Z): the queue items rewritten as `lease pool 48 --label ... --owner adv-cache-2 -- adv-cache-2 ... --threads {cores}` (16 for the plant, 32 for the windows census), a lease-aware drain in place of the flock drain. Box 1: the weak-day plant (done, fired), the 16,384-day scan (started on 32 cores), the 2^32 line census, the 4,096 drawn-era windows. Box 2: the Devnet 3 v3 diagnostic at 2^24 (waiting, 0 cores free at 19:23:53Z), the drawn census resumed from program 12 (the harness takes a program range, so nothing re-runs), the 32 drawn-era programs, devnet 2^26 under v2 |