From b1e45d914347b5647115952e632d1631c51add6e Mon Sep 17 00:00:00 2001 From: igneum-labs <337424239+igneum-labs@users.noreply.github.com> Date: Wed, 7 Oct 2026 20:38:58 +0000 Subject: [PATCH] Counter ASIC 3.0 status: the founder's premium question and section 5.10's answer, the research lane's reading of the three shapes, the five ratios and the defender's ruling (the floor closes the hot-set half, not the class), adv-mixer-2's confirmation of AP-F4-1 Co-Authored-By: Claude Fable 5.1 --- docs/plans/counter-asic-3-status.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/plans/counter-asic-3-status.md b/docs/plans/counter-asic-3-status.md index 7c18c5f2..344e5ca6 100644 --- a/docs/plans/counter-asic-3-status.md +++ b/docs/plans/counter-asic-3-status.md @@ -455,7 +455,7 @@ Reading so far: the class v4 premium at the unlocked clock is +145 W on this car | 1,400 | 134.98 | 316.3 | 0.427 | 134.68 | 228.0 | 0.591 | 1,387 | | unlocked, end | 136.75 | 473.0 | | 136.5x | 329.7 | | 2,843 / 2,850 | -Reading: the class v4 premium is 145.3 W at the unlocked clock (not the 80 W of 6 October, which was read at the app's tuned cap) and 88.3 W at the 1,400 MHz lock; v4's rate is +0.18 percent over v3 unlocked and +0.23 percent at the lock; the rate is memory-bound on the whole grid (136.8 to 135.0 MH/s from 2,850 to 1,400); the best MH per watt sits at the lowest lock on the grid, so the knee is below 1,400 MHz. the founder's thesis holds in part: 57 W of the 145 W premium comes back by the lock alone; 88 W stays as the shadow's ALU work at the floor clock. Throttle reasons: the SW power-cap governor (0x400) from unlocked to 2,163, the lock itself (0x4) from 2,100 down on v4 and 1,650 down on v3; 75 C at the top, 59 C at 1,400. The owed 5090 clock rows (2,781 / 2,472 / 2,163 / 1,854) are in this table. Per tier: a 5090 owner on class v4 who locks the core at 1,400 MHz pays 316 W instead of 476 W for 1.4 percent less rate, MH per watt up 48 percent, the v4 premium down from 145 to 88 W; the lever is NVIDIA's -lgc through the helper; AMD has the helper's ADLX tune line or nothing; the Mac has no lever. MAIN'S ORDERS ON IT (20:2x UK): (1) the second pass now, 1,400 MHz down to the driver's floor in 100 MHz steps on v4 and the v3 control, to find the knee and the premium at it, then the same grid on the 5080; (2) the knob into Ember Tune for 0.3.23 (after the power-cap search, a core-clock search downward from the cap's point until the rate falls more than 1 percent, taking the best MH/W, the fingerprint on every step, stored per card; the Mac stated as no lever) through the UI lane; (3) the bench table's 5090 row gains the locked point and the class v4 cost column reads the locked premium beside the unlocked one (done: the 1,400 MHz row 134.98 MH/s at 316.3 W, 0.427 MH/W, the Hive values 1,400 / 13,801 / 575 as the driver's default limit). DEVNET 3's FIRST LOCK (the fleet lane): 19:02:46Z, checkpoint 235 LOCKED on block 50266abe... (blue score 7,050), identical on dn3-g1 (signed 98.4 percent of active) and dn3-g2 (100.1 percent), at DAA about 7,298 (the 7,200 window filled at 19:01Z); wave 1 of the second nodes GREEN on all five at 19:17:00Z, wave 2 from 19:17Z; the 0.3.22 pin candidate 34a2dbaa passed its last gate at 19:07Z (the join-and-restart read on dn3-c1 with the N15 line) and the Devnet 3 sweep to it runs. The fleet's 19:03Z table: 43 rows, 44 GPUs, USD 257 a day, today about USD 460 of the ceiling at 19:20Z; dn3-twin (a broken CUDA host) and dn3-q02 (never answered) destroyed; p12-vast, w-target and w-poison repurposed to Devnet 3. THE SITE: a0e0c83a deployed 19:17:28Z with the Hive column. THE BOXES: build-1 read 601 / 552 / 471 and build-2 401 / 417 / 400 at 19:17Z, the sum being adversarial binaries started by hand over ssh at 64 to 89 threads each (the crypto lane's one-sweep lock not holding the sum); MAIN'S RULE for every lane under this one: no run starts on a box except through the build-server lane's `lease pool -- cmd` (landing within the quarter hour); hand-started runs killed by their kill files and re-queued through the lease; the release builds and the v5 suites outrank the sweeps. THE SHIPPER cuts class v5 as 0.3.24 the minute every v5 gate is green, at any hour; this lane's gate board is the only clock. THE (c''') CENSUS NUMBER (the v5 lane, 21:03 UK; 4,600 f8 seeds, box 2, 48 cores, 1,256 s; log docs/design/class-v5-harness/v5-census-4600-0.log): the 0.995 per-site floor rejects 112 of 4,600 class v4 sub-version 3 accepted programs (2.435 percent); the same 112 move to a later class v5 attempt; the attempts mean 2.174 to 2.248 (+3.4 percent; the tail unchanged, max 24 on both); 0 class v5 accepted programs under the floor. The spread of the v4-accepted programs' minimum site ratio at the 2^20 sample: min 0.9807, p0.1 0.9831, p1 0.9906, p5 0.9962, median 0.9999; under 0.98 none, under 0.99 43 (0.935 percent), under 0.995 112, under 0.998 517, under 0.999 868. Two corrections to the relayed premise: the clean spread is not "0.9960 minimum, p1 0.9990" on the chain's own draw (a smaller sample), and 0.99 would NOT refuse the exemplar (0.9919). So 0.995 is the lowest round floor that refuses seed 100767 with the model's spread under it, at one extra draw attempt on 2.4 percent of epochs (about 0.3 s of acceptance each, no consensus cost, no change to the hash or the kits); the band it rejects is where every measured program so far is a weak hot-set program (adv-accept's live-low20 row: the first read, seed 3664 at 1.31x, beyond the 1.2x gate; its four measured lowest-ratio seeds all beyond). The per-site hot-item test is the same statistic at the 2^20 sample and costs 30 s per candidate at 2^24 against the floor's 0 extra, so it is not a competitor. The defender's call: keep 0.995; the freeze commit carries the number into accept.rs and section 14 and goes the minute the full suite and the gate read green (f17849eb staged). THE PRE-PUBLIC SCRUB: master's text pass (9b8eb23a, 3b4b6c63) names the founder as "the founder" in every tracked text file and the CI check founder-strings refuses the name; this record follows it from here (three lines of the efficiency pass reintroduced the name through a merge and are fixed). THE V5 FAST-TIME LINES ON THE FIRST MATCHED PAIR 959b57c9 (igneumd 519ee6c4..., igneum-pow ab6f980b; harness v5-fasttime 6de9cf74): (1) the ladder climb plus the v5 crossing with the stale node: rung 1 by miner signal from epoch 6 at 19:56:38Z; class v5 by signal from epoch 8 (DAA 480) at rung 1 at 19:58:31Z on 3 of 3 honest nodes, byte 6 at 9,985 bps, before the floor; the prelude reads v4 at epochs 0 and 1 (b1680b57 verified); honest nodes 0 PoW rejections; the stale node 122 of 122 refused by its own node; the SUMMARY FAIL at 20:04:00Z on line 3 alone. (2) Digest-compat SUMMARY PASS 20:08:30Z (one digest c0d6998e with the v5 key absent on the new binary and on the pre-v5 control 2720d8d2; the set key moves it to 8d8f6208 with no peer and the refusal line; 180 s of mixed mining, 79 new and 101 old blocks, 0 rejected). (3) THE RESTART STEP FAILED: n2 stopped at DAA 457 and restarted on its own datadir at DAA 500 (19:59:10Z), the IBD catch-up engaged and did not complete: the relay flow treated the engine's "class v5 needs the execution state" on a relayed epoch-8 block as a rule refusal twelve seconds in, the flow error tore down the peer's flows and the catch-up's body sync with them, the block went into the N6 refused memory, the peer was banned 600 s at 20:00:09Z, and the deferred headers were never validated; n2 on its own fork to the end (600 against 660). FIXED by the node lane at a3b2049d on class-v5-node-wire (20:0xZ): the relay flow holds off a block in the v5 state-wait class at its three validation sites (no strike, no N6 memory, the flow stays up; the next inv or the catch-up's deferred chunk brings it back once the executor has the state), is_v5_state_wait shared with the IBD catch-up, kaspa-p2p-flows 38 of 38 at 20:08:23Z. (4) One program id per epoch equal to the CLI's v5 id: 11 of 11 (v4 at rungs 0 and 1, v5 at rung 1 with the window's stream), the Devnet 3 genesis pack's e5a4ac5978462156 reproduced; the known-failed shape FAIL as it must at 19:04:01Z. THE SECOND MATCHED PAIR 63524e28 landed 20:16:49Z on build-1 (/srv/artefacts/v5-pair-63524e28/, igneumd sha256 4f93697f5b97c30205b7e1c8e521924ba85f849a5d35f1470ff0a1b2a17cbbae; fork v5-object-0323 = 959b57c9 + the relay hold-off a3b2049d + the pool lane's tag rename 7455b8d5 + main's chain id by height, Devnet 3 chain id 4464 from the v5 floor; igneum-pow ab6f980b; suites consensus 126, core 154, pow 19, exec 46, p2p-flows 38, miner 25); the fast-time lane re-runs the crossing with the restart step on it; the v5 object commit waits on that PASS, the freeze commit and dn3-g1's DAA at the cut. THE IN-HOUSE PASS: adv-accept's tally at 2^24: of the 16 lowest stand-in-ratio seeds, 9 beyond the 1.2x gate and 3 hot sets (100767 at 2.05x; 4346 at X/f 1.78, 2.24x; 5245 at 1.29, 1.86x); of 17 random accepted programs 1 beyond and 0 hot sets; the selector's false-positive rate 7 of 16 on the gate, 13 of 16 on the hot-set test; each hot set about 1 MB of items at 0.3 percent of reads, 1.002x; whether the 0.995 floor refuses 4346 and 5245 is a measurement in hand (their minimum-site ratios at 2^20), not a given. adv-accept-2 at its natural end (92168536): header grinding BOUND by measurement and by tail (1e8 hashes per real program on the windowed random baseline down to the 3e-7 tail, net gain 3e-7x); the rotate-identity repeat class present in 21.0 percent of 300 drawn programs at under 0.1 percent of loads on the worst, absent from the two real programs, a rule question for the next class. Lanes at a natural end: adv-cache, adv-accept-2. THE SITE: b8f501e9 deployed 20:17:42Z with the compact table; its 1600 px capture read five columns clipped at the article column's edge (about 900 px against a 1,400 px table), so the table now fits the column (nine columns, the generator in the detail row, the text columns wrapping and the numbers not, no forced width). A SHARED-DEVNET FACT FROM THE FLEET (not this lane's, with the shipper and the infra lane): the Hetzner live seed 188.245.5.161:26611 is still on the old override object (digest eada4bda) 1 h 40 min after the 0.3.20 sweep (the fleet never touches Hetzner nodes, so it was outside the sweep); the 0.3.21 wipe canary c22-1 took five digest-mismatch rejects from it; an app with the packaged peers is refused at the seed and syncs through node1 and the hub only, a fresh joiner with only the seed cannot join, the 14 voters and the hub are unaffected; the owner puts the floor file ov16-floor-900000.json (sha 294f1f80) and the c4459193 pin on it. 0.3.21's STAGING (the node lane): the order dry-merges onto 55768f88 with nothing moving to 0.3.22; the late-join fix is 52e96c94 (70e4601e rebased onto 55768f88, exec suite 33 green with both new tests); f067f7c1, b0444f51 and 437f0438 merge clean in order; 2e32d5f6's one conflict (DST_ADDRESS beside pool-finish's DST_BINDING in consensus/core/src/finality.rs) kept both; the live-file digest eada4bda after each (every switch at never); the staging waits on the shipper's sweep-end word; the re-pin held. PC 2 DOWN AGAIN (main, 16:5x UK): the founder takes PC 2 down for cable work (PC 1 back but his desk); both PCs out of the sweep's waves, each updates on its poller on return; no PC job to PC 1; the Windows G1 completed before the outage, nothing reruns. 0.3.21's SECOND GATE LINE on 55768f88 (sha256 279b1b690e854fc9): the ten-minute mixed-version gate beside the 5899f603 pair, 13:37:40Z to 13:47:52Z, SUMMARY PASS (one digest b0afb2ee on five nodes; 223 new and 381 old blocks accepted by the old hub, 0 rejected; counts equal at 319, 486 and 604 through both clean joins and the restart step at 13:45:22Z; no panic); the node lane's two lines on 0.3.21's first candidate complete, in plan 6.9 on ca3-v4-node; the fleet's set on it (the bare-child 12 GB line, the wipe, the kept read, the cases) is the fleet's. 0.3.21's FIRST GATE LINE on 55768f88 (sha256 279b1b690e854fc9, the string read back; pairing igneum-pow 8c728ca3 at byte 5): the digest gate 13:35:41Z to 13:37:19Z SUMMARY PASS (a89be8a7 on both binaries with the peers; db9a85f9 refused, no peer; the live file's eada4bda unmoved); the ten-minute mixed-version gate from 13:37:40Z, line about 13:50Z. The 0.3.21 order as the shipper sent it: 55768f88; f067f7c1 and 70e4601e; b0444f51; 6eb21fc9; db28d331; then the re-pin from 8bdcbdd8 on the coordinator's word; suites between, the digest read after every one; the mirror's release-0.3.20-node back at the pin c4459193, release-0.3.21-node open at 55768f88. THE LATE-JOIN COMMIT (N9's second half, the node lane): 70e4601e on the box mirror as branch proof-hold-fix, from c4459193, two files (igneum/exec/src/proving.rs, protocol/flows/src/v10/proving.rs); the gap was the fetch side on the joiner (the served record ran the native check against the joiner's trailing exec state before anything was stored, the check refused it, the proof was never held, the body rule read "not held" for 20 s and failed the IBD); the fix holds the proof by hash before the checks (the pool entry still needs them) and the serve side says when it holds fewer than asked; the exec suite 32 passed at 13:26Z with the known-failed shape first, the flows check green 13:28Z, igneumd on build-1 at the 0321 worktree path built 13:32Z, sha256 17649eeb2f7d1290, string read back; with the testnet lane (the resume form, B alone); it joins the 0.3.21 staging as its own commit. THE WIPE CANARY ON c19-1, c4459193 (sha 45be9b02d1b002f5, string read back): FORM END rc 0 at 13:50:53Z. Wipe synced 13:35:50Z (57 minutes, inside the 98-minute class); mining 13:36:00Z to 13:47:07Z, 66 mined, 66 accepted, 0 rejected, isSynced true at the tip throughout; the hub holds 41 of its blocks in its last 700 with 0 rejects (13:47:09Z); the restart on its kept datadir at 13:47:15Z: the old process stopped at once (the new process's first lock line seven seconds after the marker; the watchdog held nothing, the b7cc37e7 fault closed), synced again at 13:48:39Z after 84 s, 109 templates read with max 3,432 ms and 0 timeouts; the kept read on pool-1's 0.3.17 copy on the same pod passed at 13:38Z (the rewrite line once, a clean second start). The pin's set on c4459193: the digest gate PASS, the mixed-version gate PASS, the wipe canary PASS, the kept read PASS, the restart PASS, the 12 GB line proves and verifies (paid is a race, not a gate); CASES END from c20-1 (about 14:50Z) is the last pin line. THE INTEROP FACT stands from the void run: the 5899f603 hub accepted 235 object-byte-5 blocks from the 8097d600 node with 0 rejected, one digest on all five nodes on the live sixteen-field file. The gates: the digest test and the kaspa-pow vector test (the amended devnet epoch-0 id 1a4230699a6b9c60 must equal, c120d7963abdcd96 must differ, the v3 control unchanged) on the box; the mixed-version Devnet 2 gate (the amended 0.3.20 node beside a 5899f603 node for ten minutes on the live file without the v4 fields) after the Mac build; the fresh-join canary the 0.3.20 cut's | +Reading: the class v4 premium is 145.3 W at the unlocked clock (not the 80 W of 6 October, which was read at the app's tuned cap) and 88.3 W at the 1,400 MHz lock; v4's rate is +0.18 percent over v3 unlocked and +0.23 percent at the lock; the rate is memory-bound on the whole grid (136.8 to 135.0 MH/s from 2,850 to 1,400); the best MH per watt sits at the lowest lock on the grid, so the knee is below 1,400 MHz. the founder's thesis holds in part: 57 W of the 145 W premium comes back by the lock alone; 88 W stays as the shadow's ALU work at the floor clock. Throttle reasons: the SW power-cap governor (0x400) from unlocked to 2,163, the lock itself (0x4) from 2,100 down on v4 and 1,650 down on v3; 75 C at the top, 59 C at 1,400. The owed 5090 clock rows (2,781 / 2,472 / 2,163 / 1,854) are in this table. Per tier: a 5090 owner on class v4 who locks the core at 1,400 MHz pays 316 W instead of 476 W for 1.4 percent less rate, MH per watt up 48 percent, the v4 premium down from 145 to 88 W; the lever is NVIDIA's -lgc through the helper; AMD has the helper's ADLX tune line or nothing; the Mac has no lever. MAIN'S ORDERS ON IT (20:2x UK): (1) the second pass now, 1,400 MHz down to the driver's floor in 100 MHz steps on v4 and the v3 control, to find the knee and the premium at it, then the same grid on the 5080; (2) the knob into Ember Tune for 0.3.23 (after the power-cap search, a core-clock search downward from the cap's point until the rate falls more than 1 percent, taking the best MH/W, the fingerprint on every step, stored per card; the Mac stated as no lever) through the UI lane; (3) the bench table's 5090 row gains the locked point and the class v4 cost column reads the locked premium beside the unlocked one (done: the 1,400 MHz row 134.98 MH/s at 316.3 W, 0.427 MH/W, the Hive values 1,400 / 13,801 / 575 as the driver's default limit). DEVNET 3's FIRST LOCK (the fleet lane): 19:02:46Z, checkpoint 235 LOCKED on block 50266abe... (blue score 7,050), identical on dn3-g1 (signed 98.4 percent of active) and dn3-g2 (100.1 percent), at DAA about 7,298 (the 7,200 window filled at 19:01Z); wave 1 of the second nodes GREEN on all five at 19:17:00Z, wave 2 from 19:17Z; the 0.3.22 pin candidate 34a2dbaa passed its last gate at 19:07Z (the join-and-restart read on dn3-c1 with the N15 line) and the Devnet 3 sweep to it runs. The fleet's 19:03Z table: 43 rows, 44 GPUs, USD 257 a day, today about USD 460 of the ceiling at 19:20Z; dn3-twin (a broken CUDA host) and dn3-q02 (never answered) destroyed; p12-vast, w-target and w-poison repurposed to Devnet 3. THE SITE: a0e0c83a deployed 19:17:28Z with the Hive column. THE BOXES: build-1 read 601 / 552 / 471 and build-2 401 / 417 / 400 at 19:17Z, the sum being adversarial binaries started by hand over ssh at 64 to 89 threads each (the crypto lane's one-sweep lock not holding the sum); MAIN'S RULE for every lane under this one: no run starts on a box except through the build-server lane's `lease pool -- cmd` (landing within the quarter hour); hand-started runs killed by their kill files and re-queued through the lease; the release builds and the v5 suites outrank the sweeps. THE SHIPPER cuts class v5 as 0.3.24 the minute every v5 gate is green, at any hour; this lane's gate board is the only clock. THE (c''') CENSUS NUMBER (the v5 lane, 21:03 UK; 4,600 f8 seeds, box 2, 48 cores, 1,256 s; log docs/design/class-v5-harness/v5-census-4600-0.log): the 0.995 per-site floor rejects 112 of 4,600 class v4 sub-version 3 accepted programs (2.435 percent); the same 112 move to a later class v5 attempt; the attempts mean 2.174 to 2.248 (+3.4 percent; the tail unchanged, max 24 on both); 0 class v5 accepted programs under the floor. The spread of the v4-accepted programs' minimum site ratio at the 2^20 sample: min 0.9807, p0.1 0.9831, p1 0.9906, p5 0.9962, median 0.9999; under 0.98 none, under 0.99 43 (0.935 percent), under 0.995 112, under 0.998 517, under 0.999 868. Two corrections to the relayed premise: the clean spread is not "0.9960 minimum, p1 0.9990" on the chain's own draw (a smaller sample), and 0.99 would NOT refuse the exemplar (0.9919). So 0.995 is the lowest round floor that refuses seed 100767 with the model's spread under it, at one extra draw attempt on 2.4 percent of epochs (about 0.3 s of acceptance each, no consensus cost, no change to the hash or the kits); the band it rejects is where every measured program so far is a weak hot-set program (adv-accept's live-low20 row: the first read, seed 3664 at 1.31x, beyond the 1.2x gate; its four measured lowest-ratio seeds all beyond). The per-site hot-item test is the same statistic at the 2^20 sample and costs 30 s per candidate at 2^24 against the floor's 0 extra, so it is not a competitor. The defender's call: keep 0.995; the freeze commit carries the number into accept.rs and section 14 and goes the minute the full suite and the gate read green (f17849eb staged). THE PRE-PUBLIC SCRUB: master's text pass (9b8eb23a, 3b4b6c63) names the founder as "the founder" in every tracked text file and the CI check founder-strings refuses the name; this record follows it from here (three lines of the efficiency pass reintroduced the name through a merge and are fixed). THE V5 FAST-TIME LINES ON THE FIRST MATCHED PAIR 959b57c9 (igneumd 519ee6c4..., igneum-pow ab6f980b; harness v5-fasttime 6de9cf74): (1) the ladder climb plus the v5 crossing with the stale node: rung 1 by miner signal from epoch 6 at 19:56:38Z; class v5 by signal from epoch 8 (DAA 480) at rung 1 at 19:58:31Z on 3 of 3 honest nodes, byte 6 at 9,985 bps, before the floor; the prelude reads v4 at epochs 0 and 1 (b1680b57 verified); honest nodes 0 PoW rejections; the stale node 122 of 122 refused by its own node; the SUMMARY FAIL at 20:04:00Z on line 3 alone. (2) Digest-compat SUMMARY PASS 20:08:30Z (one digest c0d6998e with the v5 key absent on the new binary and on the pre-v5 control 2720d8d2; the set key moves it to 8d8f6208 with no peer and the refusal line; 180 s of mixed mining, 79 new and 101 old blocks, 0 rejected). (3) THE RESTART STEP FAILED: n2 stopped at DAA 457 and restarted on its own datadir at DAA 500 (19:59:10Z), the IBD catch-up engaged and did not complete: the relay flow treated the engine's "class v5 needs the execution state" on a relayed epoch-8 block as a rule refusal twelve seconds in, the flow error tore down the peer's flows and the catch-up's body sync with them, the block went into the N6 refused memory, the peer was banned 600 s at 20:00:09Z, and the deferred headers were never validated; n2 on its own fork to the end (600 against 660). FIXED by the node lane at a3b2049d on class-v5-node-wire (20:0xZ): the relay flow holds off a block in the v5 state-wait class at its three validation sites (no strike, no N6 memory, the flow stays up; the next inv or the catch-up's deferred chunk brings it back once the executor has the state), is_v5_state_wait shared with the IBD catch-up, kaspa-p2p-flows 38 of 38 at 20:08:23Z. (4) One program id per epoch equal to the CLI's v5 id: 11 of 11 (v4 at rungs 0 and 1, v5 at rung 1 with the window's stream), the Devnet 3 genesis pack's e5a4ac5978462156 reproduced; the known-failed shape FAIL as it must at 19:04:01Z. THE SECOND MATCHED PAIR 63524e28 landed 20:16:49Z on build-1 (/srv/artefacts/v5-pair-63524e28/, igneumd sha256 4f93697f5b97c30205b7e1c8e521924ba85f849a5d35f1470ff0a1b2a17cbbae; fork v5-object-0323 = 959b57c9 + the relay hold-off a3b2049d + the pool lane's tag rename 7455b8d5 + main's chain id by height, Devnet 3 chain id 4464 from the v5 floor; igneum-pow ab6f980b; suites consensus 126, core 154, pow 19, exec 46, p2p-flows 38, miner 25); the fast-time lane re-runs the crossing with the restart step on it; the v5 object commit waits on that PASS, the freeze commit and dn3-g1's DAA at the cut. THE IN-HOUSE PASS: adv-accept's tally at 2^24: of the 16 lowest stand-in-ratio seeds, 9 beyond the 1.2x gate and 3 hot sets (100767 at 2.05x; 4346 at X/f 1.78, 2.24x; 5245 at 1.29, 1.86x); of 17 random accepted programs 1 beyond and 0 hot sets; the selector's false-positive rate 7 of 16 on the gate, 13 of 16 on the hot-set test; each hot set about 1 MB of items at 0.3 percent of reads, 1.002x; whether the 0.995 floor refuses 4346 and 5245 is a measurement in hand (their minimum-site ratios at 2^20), not a given. adv-accept-2 at its natural end (92168536): header grinding BOUND by measurement and by tail (1e8 hashes per real program on the windowed random baseline down to the 3e-7 tail, net gain 3e-7x); the rotate-identity repeat class present in 21.0 percent of 300 drawn programs at under 0.1 percent of loads on the worst, absent from the two real programs, a rule question for the next class. Lanes at a natural end: adv-cache, adv-accept-2. THE SITE: b8f501e9 deployed 20:17:42Z with the compact table; its 1600 px capture read five columns clipped at the article column's edge (about 900 px against a 1,400 px table), so the table now fits the column (nine columns, the generator in the detail row, the text columns wrapping and the numbers not, no forced width). THE FOUNDER'S QUESTION ON THE CLASS V4 PREMIUM ("we need a solution, deep research, other methods, something must be doable even if it is revolutionary"; a research lane, counter-asic-4, on the literature and the alternatives by 23:30Z). THE ONE COMPUTATION FROM THIS LANE (chip-model-v3 section 5.10, on the mirror's master at aa829826): class v5 ON and the shadow at ZERO leaves the strongest chip, the f = 1 stored-dataset chip, at 5.1x (GDDR7, the 5090's own 16 devices without the GPU: 166 MH/s at 78 W with a farm-shared node) to 9.1x (HBM3 eight stacks: 666 MH/s at 175 W) per joule over the 5090's 0.417 MH/W, the class v3 figures of 5.6 less a rounding, because the node is a farm cost and not a chip cost (class-v5 2a.2: one node serves a farm, the leaves ship at 16.5 KB/s to 10,000 members, the rebuild is the same 32 ms per window every GPU pays); only a chip forced to carry its own node (85 W, approximate) falls near 2x, and only the small ones (one HBM3 stack 1.8x, the GDDR7 board 2.5x; the eight-stack package 6.2x). So the shadow stays the only lever in the model that reaches the memory-system chip, and the premium is its price; what class v5 buys is the recompute and stale chips gone as categories and every miner holding and following the chain. THE RESEARCH LANE's READING of the three shapes (21:4x UK; its identity edge = (E_card + F) / (E_mem + k F), F the GPU's premium per hash, k the chip core's energy per op over the GPU's): at the 1,400 MHz lock (v3 1.69 uJ per hash, v4 2.34, the premium 0.65 uJ = 6.5 pJ per counted op) the premium-free edge is 3.6x on GDDR7 and 5.3x on one HBM3 stack, 2.1x at F = 0.65 and k = 1, the asymptote 1/k; the premium needed for 2x at k = 1 is 0.76 uJ (103 W) at the lock and 1.41 uJ (175 W) at the stock point; at k at or under 0.5 no premium reaches 2x; a premium of zero is impossible by any hash-side lever (a joule the chip must spend is a joule the GPU spends first). Shape 1 (the shadow at the floor clock): correct and measured by the 1,400 row; it halves the premium and leaves k as the whole question (the edge's derivative in F at F = 0 is minus 5.7x per uJ at k = 1 and minus 0.2x per uJ at k = 0.3; the family mix with the highest k, shuffles and multiplies, is the right content; the tensor block has k near 1 but 0.056 pJ per MAC, so it forces no joules without 26x the verifier). Shape 2 (per-read work a chip cannot amortise): no construction found (the chip's lane count is set by its own latency, lane state is SRAM at pJ per access, extra reads scale both rates; row, bank, refresh and burst shaping have no asymmetry); one candidate for a measurement, not a claim: a hot table kept L2-resident through cache-policy hints (dataset loads evict-first, hot loads evict-last), since a 64 MiB SRAM read on a chip is not cheaper than a GPU L2 hit (the 3.0 layer 5 measurement used no hints and lost 13 to 16 percent). Shape 3 (the refresh as the cost): dead by arithmetic (a 1 GiB rebuild about 0.1 J on a chip against 280 kJ of hashing an hour). What lowers the premium-free floor is the card's own E_card: the 5090's idle 74 to 91 W plus its memory system's 55 W bound the floor near 2.0x on GDDR7 at any operating point; the two measurements that read how much of the 100 W between is reachable are the knee below 1,400 MHz (the floor pass, on PC 1's queue) and an SM-sparse kernel (the hash on a fraction of the SMs with 4 to 8 chains per thread, the rest clock-gated), both ordered to the hash lane with the L2 cache-policy hot table as PC 1 queue tail items (worker launch shapes, --bench only, no consensus change). THE FIVE RATIOS (adv-accept, 21:35 BST, at the acceptance's 2^20 sample, closed form and live to 1e-4): 4346 (bbb38e847011c354) minimum site 2 at 0.9840 REFUSED by (c''') at 0.995; 5245 (beaad44840bb9e4e) site 8 at 0.9831 REFUSED; 106924 at 0.9821 and 107022 at 0.9877 REFUSED; Devnet 3's epoch-0 class v4 program (fce15bf61030be57) site 0 at 0.9992 (live 0.9993) MISSED; era-fixed-20 (11f9f955b21d56c9) site 11 at 0.9997 MISSED; era-drawn-2 (7ceb797d31eedb3e) site 13 at 0.9992 MISSED. THE DEFENDER'S RULING: the (c''') floor at 0.995 closes the HOT-SET HALF of the class (every program with X_f at or above f that any lane measured, 100767 included) and NOT the class; the residual is the shadow-block-written value-level concentration below a ratio floor's resolution (1.26x to 1.45x of the top 0.1 percent share, 0.03 to 0.1 percent of a hash's reads each, minimum sites 0.9992 to 0.9997, which no floor reaches without sitting inside the clean seeds' own spread), chip gain under 1.001x, its lever a value-level source test at live scale or a per-site hot-item test, routed to the next class as a named item; section 14 of the v5 design says so with the five numbers verbatim; the freeze proceeds on the suite's green, both the floor's number and its reach measured. adv-mixer-2 (49656c2e) confirms AP-F4-1's class from outsider inputs (model A an FPGA LUT-area gain for a per-day build, 1.0 on every GPU, verifier and chip with a general multiplier; P(A at or above 1.1x) = 2^-10.8 per day on the exact median 226; model C under 2^-20 at 1.1x), the two censuses' medians (226 against 231) to be reconciled side by side in the record; its redraw rule to the v5 lane as a second independent statement. The kits lane's box-side body moved to tools/class-v5/kits-on-box.sh run by path under the inline-rm rule (v5-kits 4f9d96d2); the kit zip stands. A SHARED-DEVNET FACT FROM THE FLEET (not this lane's, with the shipper and the infra lane): the Hetzner live seed 188.245.5.161:26611 is still on the old override object (digest eada4bda) 1 h 40 min after the 0.3.20 sweep (the fleet never touches Hetzner nodes, so it was outside the sweep); the 0.3.21 wipe canary c22-1 took five digest-mismatch rejects from it; an app with the packaged peers is refused at the seed and syncs through node1 and the hub only, a fresh joiner with only the seed cannot join, the 14 voters and the hub are unaffected; the owner puts the floor file ov16-floor-900000.json (sha 294f1f80) and the c4459193 pin on it. 0.3.21's STAGING (the node lane): the order dry-merges onto 55768f88 with nothing moving to 0.3.22; the late-join fix is 52e96c94 (70e4601e rebased onto 55768f88, exec suite 33 green with both new tests); f067f7c1, b0444f51 and 437f0438 merge clean in order; 2e32d5f6's one conflict (DST_ADDRESS beside pool-finish's DST_BINDING in consensus/core/src/finality.rs) kept both; the live-file digest eada4bda after each (every switch at never); the staging waits on the shipper's sweep-end word; the re-pin held. PC 2 DOWN AGAIN (main, 16:5x UK): the founder takes PC 2 down for cable work (PC 1 back but his desk); both PCs out of the sweep's waves, each updates on its poller on return; no PC job to PC 1; the Windows G1 completed before the outage, nothing reruns. 0.3.21's SECOND GATE LINE on 55768f88 (sha256 279b1b690e854fc9): the ten-minute mixed-version gate beside the 5899f603 pair, 13:37:40Z to 13:47:52Z, SUMMARY PASS (one digest b0afb2ee on five nodes; 223 new and 381 old blocks accepted by the old hub, 0 rejected; counts equal at 319, 486 and 604 through both clean joins and the restart step at 13:45:22Z; no panic); the node lane's two lines on 0.3.21's first candidate complete, in plan 6.9 on ca3-v4-node; the fleet's set on it (the bare-child 12 GB line, the wipe, the kept read, the cases) is the fleet's. 0.3.21's FIRST GATE LINE on 55768f88 (sha256 279b1b690e854fc9, the string read back; pairing igneum-pow 8c728ca3 at byte 5): the digest gate 13:35:41Z to 13:37:19Z SUMMARY PASS (a89be8a7 on both binaries with the peers; db9a85f9 refused, no peer; the live file's eada4bda unmoved); the ten-minute mixed-version gate from 13:37:40Z, line about 13:50Z. The 0.3.21 order as the shipper sent it: 55768f88; f067f7c1 and 70e4601e; b0444f51; 6eb21fc9; db28d331; then the re-pin from 8bdcbdd8 on the coordinator's word; suites between, the digest read after every one; the mirror's release-0.3.20-node back at the pin c4459193, release-0.3.21-node open at 55768f88. THE LATE-JOIN COMMIT (N9's second half, the node lane): 70e4601e on the box mirror as branch proof-hold-fix, from c4459193, two files (igneum/exec/src/proving.rs, protocol/flows/src/v10/proving.rs); the gap was the fetch side on the joiner (the served record ran the native check against the joiner's trailing exec state before anything was stored, the check refused it, the proof was never held, the body rule read "not held" for 20 s and failed the IBD); the fix holds the proof by hash before the checks (the pool entry still needs them) and the serve side says when it holds fewer than asked; the exec suite 32 passed at 13:26Z with the known-failed shape first, the flows check green 13:28Z, igneumd on build-1 at the 0321 worktree path built 13:32Z, sha256 17649eeb2f7d1290, string read back; with the testnet lane (the resume form, B alone); it joins the 0.3.21 staging as its own commit. THE WIPE CANARY ON c19-1, c4459193 (sha 45be9b02d1b002f5, string read back): FORM END rc 0 at 13:50:53Z. Wipe synced 13:35:50Z (57 minutes, inside the 98-minute class); mining 13:36:00Z to 13:47:07Z, 66 mined, 66 accepted, 0 rejected, isSynced true at the tip throughout; the hub holds 41 of its blocks in its last 700 with 0 rejects (13:47:09Z); the restart on its kept datadir at 13:47:15Z: the old process stopped at once (the new process's first lock line seven seconds after the marker; the watchdog held nothing, the b7cc37e7 fault closed), synced again at 13:48:39Z after 84 s, 109 templates read with max 3,432 ms and 0 timeouts; the kept read on pool-1's 0.3.17 copy on the same pod passed at 13:38Z (the rewrite line once, a clean second start). The pin's set on c4459193: the digest gate PASS, the mixed-version gate PASS, the wipe canary PASS, the kept read PASS, the restart PASS, the 12 GB line proves and verifies (paid is a race, not a gate); CASES END from c20-1 (about 14:50Z) is the last pin line. THE INTEROP FACT stands from the void run: the 5899f603 hub accepted 235 object-byte-5 blocks from the 8097d600 node with 0 rejected, one digest on all five nodes on the live sixteen-field file. The gates: the digest test and the kaspa-pow vector test (the amended devnet epoch-0 id 1a4230699a6b9c60 must equal, c120d7963abdcd96 must differ, the v3 control unchanged) on the box; the mixed-version Devnet 2 gate (the amended 0.3.20 node beside a 5899f603 node for ten minutes on the live file without the v4 fields) after the Mac build; the fresh-join canary the 0.3.20 cut's | | Main's rulings (7 October, morning) | no generator change to v4 on the live devnet; the record's null is the window model with numbers, sent by the hash lane to the attack-pass lane so AP-F8-1 re-gates against it; a fault beyond the model (a low-entropy source at site 15) stops at the coordinator with the two options priced (a 0.3.19 class amendment before the flip, or the flip held at the floor), nothing shipping without the founder's word; the tighter tail, an acceptance bound on the hot-set share, is a CLASS V5 item (sent to the v5 lane a6410f3b8abefb762 with the 64-seed census as its gate; the bound's number follows from the model) | ### AP-F4-1, the weak-day MUL draw (the attack-pass lane, 7 October, morning): PASS against v4, a class v5 rule