Embedded pairs and step-size control, measured off-archive.
The rk run ended on 2026-11-10, and no more compute goes to it. Epoch 2 is frozen and these pages are its final snapshot: the site is no longer rebuilt. The overview site sums up the final findings, and the harness is public for anyone to download and run.
An adaptive method uses an embedded pair to estimate each step's error, and a step-size controller retries any step that misses the tolerance. A record holds one set of weights and a fixed step count, so the run searches pairs in a lane of its own: cycles set aside for this class that write an unscored archive. The lane reports the modeled cycles a method needs to reach an accuracy target.
Source: side-track job adaptive.suite_sweep, 8 points. Arithmetic: float64 only; no Q15 effects are included.
| problem | tolerance | rhs evaluations | achieved error | rejected steps |
|---|---|---|---|---|
| battery_2rc | 0.001 | 28 | 0.00343958 | 0 |
| battery_2rc | 0.0001 | 37 | 0.000673436 | 0 |
| battery_2rc | 1e-05 | 55 | 0.000100945 | 0 |
| battery_2rc | 1e-06 | 97 | 1.22824e-05 | 0 |
| battery_2rc | 1e-07 | 199 | 1.44501e-06 | 1 |
| battery_2rc | 1e-08 | 421 | 1.52719e-07 | 2 |
| bicycle_lateral | 0.001 | 43 | 0.000357318 | 0 |
| bicycle_lateral | 0.0001 | 67 | 3.03138e-05 | 0 |
| bicycle_lateral | 1e-05 | 139 | 4.76471e-06 | 2 |
| bicycle_lateral | 1e-06 | 274 | 5.85663e-07 | 2 |
| bicycle_lateral | 1e-07 | 571 | 1.08566e-07 | 3 |
| bicycle_lateral | 1e-08 | 1195 | 1.50442e-08 | 3 |
| buck_converter | 0.001 | 91 | 0.00151497 | 0 |
| buck_converter | 0.0001 | 181 | 0.000447659 | 1 |
| buck_converter | 1e-05 | 376 | 7.38994e-05 | 2 |
| buck_converter | 1e-06 | 787 | 8.24192e-06 | 3 |
| buck_converter | 1e-07 | 1666 | 8.87062e-07 | 3 |
| buck_converter | 1e-08 | 3571 | 9.10891e-08 | 4 |
| enzyme_qssa | 0.0001 | 1162 | 1.7963e-05 | 2 |
| enzyme_qssa | 1e-05 | 1165 | 4.17984e-07 | 2 |
| enzyme_qssa | 0.001 | 1165 | 0.000638856 | 3 |
| enzyme_qssa | 1e-06 | 1645 | 8.16411e-07 | 109 |
| enzyme_qssa | 1e-07 | 1957 | 1.67823e-08 | 8 |
| enzyme_qssa | 1e-08 | 3301 | 2.38565e-09 | 5 |
| glucose_minimal | 0.001 | 49 | 0.000325491 | 0 |
| glucose_minimal | 0.0001 | 88 | 4.03546e-05 | 0 |
| glucose_minimal | 1e-05 | 181 | 5.1956e-06 | 2 |
| glucose_minimal | 1e-06 | 373 | 5.34465e-07 | 2 |
| glucose_minimal | 1e-07 | 787 | 5.58226e-08 | 3 |
| glucose_minimal | 1e-08 | 1675 | 5.6319e-09 | 3 |
| pll_lock | 0.001 | 52 | 0.000746124 | 0 |
| pll_lock | 0.0001 | 85 | 0.000145968 | 0 |
| pll_lock | 1e-05 | 175 | 2.60402e-05 | 2 |
| pll_lock | 1e-06 | 358 | 3.16661e-06 | 2 |
| pll_lock | 1e-07 | 757 | 3.52712e-07 | 3 |
| pll_lock | 1e-08 | 1603 | 3.67843e-08 | 3 |
| robertson_scaled | 0.001 | 2353 | 0.000597569 | 2 |
| robertson_scaled | 0.0001 | 2380 | 3.26777e-05 | 6 |
| robertson_scaled | 1e-05 | 2425 | 8.16161e-06 | 7 |
| robertson_scaled | 1e-06 | 2536 | 5.66517e-07 | 7 |
| robertson_scaled | 1e-07 | 2857 | 8.94056e-08 | 9 |
| robertson_scaled | 1e-08 | 3883 | 3.81611e-09 | 55 |
| servo_load_step | 0.001 | 3004 | 0.00187753 | 2 |
| servo_load_step | 0.0001 | 3007 | 0.000138677 | 3 |
| servo_load_step | 1e-05 | 3016 | 3.17866e-05 | 4 |
| servo_load_step | 1e-06 | 3028 | 2.84762e-06 | 4 |
| servo_load_step | 1e-07 | 3064 | 4.45106e-08 | 5 |
| servo_load_step | 1e-08 | 3145 | 6.00325e-09 | 7 |
Source: side-track job adaptive.controller_gains, 16 points. Arithmetic: float64 only; no Q15 effects are included.
| alpha | beta | rejected share | rejected steps | accepted steps | rhs evaluations |
|---|---|---|---|---|---|
| 1/8 | 0 | 0.02534 | 13 | 500 | 1542 |
| 1/8 | 1/16 | 0.01793 | 13 | 712 | 2178 |
| 1/8 | 1/8 | no run finished | n/a | n/a | n/a |
| 1/8 | 1/4 | no run finished | n/a | n/a | n/a |
| 1/4 | 0 | 0.01663 | 7 | 414 | 1266 |
| 1/4 | 1/16 | 0.01562 | 7 | 441 | 1347 |
| 1/4 | 1/8 | 0.01386 | 7 | 498 | 1518 |
| 1/4 | 1/4 | no run finished | n/a | n/a | n/a |
| 3/8 | 0 | 0.01018 | 4 | 389 | 1182 |
| 3/8 | 1/16 | 0.00993 | 4 | 399 | 1212 |
| 3/8 | 1/8 | 0.00957 | 4 | 414 | 1257 |
| 3/8 | 1/4 | 0.00797 | 4 | 498 | 1509 |
| 1/2 | 0 | 0.01309 | 5 | 377 | 1149 |
| 1/2 | 1/16 | 0.01295 | 5 | 381 | 1161 |
| 1/2 | 1/8 | 0.01269 | 5 | 389 | 1185 |
| 1/2 | 1/4 | 0.15748 | 80 | 428 | 1527 |
Not counted (space over the census cap of 100,000 matrices): 4 stages, s ≤ 2; 4 stages, s ≤ 3; 4 stages, s ≤ 4.
Source: side-track job adaptive.pair_census, 8 points. Arithmetic: exact over Fractions; no float and no Q15 arithmetic is involved.
| stages and lattice depth | condition | matrices |
|---|---|---|
| 3 stages, s ≤ 1 | matrices | 64 |
| 3 stages, s ≤ 1 | order-2 consistent | 64 |
| 3 stages, s ≤ 1 | order-3 consistent | 0 |
| 3 stages, s ≤ 1 | both solvable | 0 |
| 3 stages, s ≤ 2 | matrices | 512 |
| 3 stages, s ≤ 2 | order-2 consistent | 512 |
| 3 stages, s ≤ 2 | order-3 consistent | 4 |
| 3 stages, s ≤ 2 | both solvable | 4 |
| 3 stages, s ≤ 3 | matrices | 4,096 |
| 3 stages, s ≤ 3 | order-2 consistent | 4,096 |
| 3 stages, s ≤ 3 | order-3 consistent | 6 |
| 3 stages, s ≤ 3 | both solvable | 6 |
| 3 stages, s ≤ 4 | matrices | 32,768 |
| 3 stages, s ≤ 4 | order-2 consistent | 32,768 |
| 3 stages, s ≤ 4 | order-3 consistent | 24 |
| 3 stages, s ≤ 4 | both solvable | 24 |
| 4 stages, s ≤ 1 | matrices | 4,096 |
| 4 stages, s ≤ 1 | order-2 consistent | 4,096 |
| 4 stages, s ≤ 1 | order-3 consistent | 3,646 |
| 4 stages, s ≤ 1 | both solvable | 3,646 |
rk-work/benchmark/results.json has not been written in this work directory, so the matched-accuracy comparison is not stated here. It appears once rk_harness.benchmark has been run against the current code; nothing is inferred in its place. The epoch 1 comparison, run under that epoch's own pin, is on the frozen epoch 1 site.
The archive fixes a cycle budget and measures the error a method reaches. The lane fixes an accuracy target on the axis-T ladder and measures the modeled cycles a candidate needs to reach it, so neither number converts into the other. Every run is float64. The pinned order-condition checker never saw these coefficients, so none is order-verified.
Source: rk-work/adaptive_archive/elites.json. float64 runs, cycles priced on float_trajectory_q15_attempt_cost; not order-verified.
| rank | median cycles at the elite target | best achieved error | entries |
|---|---|---|---|
| 1 | 3388 | 9.01642e-06 | 1 |
| 2 | 3400 | 8.9257e-05 | 1 |
| 3 | 3416 | 3.77813e-05 | 1 |
| 4 | 3444 | 7.54048e-05 | 1 |
| 5 | 3472 | 2.62007e-05 | 1 |
| 6 | 3472 | 4.29037e-05 | 1 |
| 7 | 3490 | 6.27542e-05 | 1 |
| 8 | 3500 | 8.15011e-06 | 1 |
| 9 | 3528 | 5.02352e-06 | 1 |
| 10 | 3528 | 6.46039e-08 | 1 |
| 11 | 3530 | 5.87563e-05 | 1 |
| 12 | 3556 | 4.55622e-05 | 1 |
| 13 | 3567 | 7.82133e-07 | 1 |
| 17 | 3584 | 2.02167e-06 | 1 |
| 15 | 3584 | 2.51696e-05 | 1 |
| 14 | 3584 | 5.45847e-05 | 1 |
| 16 | 3584 | 7.07468e-06 | 1 |
| 18 | 3590 | 2.19819e-06 | 1 |
| 19 | 3590 | 5.34533e-05 | 1 |
| 21 | 3612 | 2.4653e-05 | 1 |
| 20 | 3612 | 7.30608e-07 | 1 |
| 22 | 3620 | 4.09451e-05 | 1 |
| 28 | 3640 | 0.000101556 | 1 |
| 23 | 3640 | 1.42391e-06 | 1 |
| 24 | 3640 | 4.73485e-05 | 1 |
| 25 | 3640 | 6.33567e-05 | 1 |
| 27 | 3640 | 6.77336e-07 | 1 |
| 26 | 3640 | 8.28288e-06 | 1 |
| 29 | 3645 | 7.09109e-05 | 1 |
| 30 | 3650 | 4.4025e-05 | 1 |
| 31 | 3660 | 0.000130083 | 1 |
| 32 | 3668 | 1.73117e-06 | 1 |
Ranking rule, quoted from the document: lowest median cycles at the elite target across the problems where the status is reached; a candidate that reached the elite target nowhere ranks below every candidate that reached it somewhere; ties broken by the count of targets reached, highest leading, then by record_hash. No clock is read and no value is sampled, so the order is total and two runs over the same records agree.
Rank is the position the document gives under the rule above. Each row shows the candidate's shape, the median modeled cycles it needed at the elite target, how many problems it reached that target on, and the cost basis its cycles were priced on.
| rank | candidate | order | embedded order | stages | safety | alpha | beta | median cycles at target | problems at target | targets reached | cost basis |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 6de647f4a17b6f7a | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3388 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 2 | 8c8104c35a37c56f | 3 | 2 | 4 | 15/16 | 3/16 | 1/16 | 3400 | 7 of 8 | 30 of 32 | float_trajectory_q15_attempt_cost |
| 3 | 2bc88eb7aca6c5d1 | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3416 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 4 | cbd48b68efc7b879 | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3444 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 5 | b4e320acf5af7bc7 | 3 | 2 | 4 | 3/4 | 3/8 | 1/8 | 3472 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 6 | 9a522fc83fb27c46 | 3 | 2 | 4 | 15/16 | 1/4 | 1/16 | 3472 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 7 | fb6aafddb0db2e06 | 3 | 2 | 4 | 7/8 | 1/8 | 0 | 3490 | 7 of 8 | 30 of 32 | float_trajectory_q15_attempt_cost |
| 8 | 30a681c43ebc6bed | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3500 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 9 | 54a0975db1e0b8e7 | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3528 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 10 | 32a60583f94dec9b | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3528 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 11 | b1794656aefd5a85 | 3 | 2 | 4 | 7/8 | 3/8 | 3/16 | 3530 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 12 | 44647068bbf43705 | 3 | 2 | 4 | 7/8 | 3/8 | 1/16 | 3556 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 13 | 9dc244a7d721c78b | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3567 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 14 | 9d937289d9603f8c | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3584 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 15 | 6ed785c91ebbede8 | 3 | 2 | 4 | 3/4 | 3/8 | 1/8 | 3584 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 16 | 5458f4da3de31874 | 3 | 2 | 4 | 15/16 | 1/4 | 1/16 | 3584 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 17 | 3e2f1a6bccd17b0a | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3584 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 18 | 47d82adcaa7d02cb | 3 | 2 | 4 | 7/8 | 3/8 | 3/16 | 3590 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 19 | 3b43487937ebcead | 3 | 2 | 4 | 7/8 | 3/8 | 3/16 | 3590 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 20 | cafe2b7a70fd5076 | 3 | 2 | 4 | 7/8 | 3/16 | 0 | 3612 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 21 | 9cd0ec9fcbe980b2 | 3 | 2 | 4 | 7/8 | 3/8 | 1/16 | 3612 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 22 | 4e063a3c7a6c1899 | 3 | 2 | 4 | 7/8 | 3/8 | 3/16 | 3620 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 23 | 11960e2d0409354b | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3640 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 24 | 650a40fe4a8ad481 | 3 | 2 | 4 | 15/16 | 1/4 | 1/16 | 3640 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 25 | 70540246b86bdedc | 3 | 2 | 4 | 3/4 | 3/8 | 1/8 | 3640 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 26 | 6af1d2d65dcc4578 | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3640 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 27 | d748277d844255f2 | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3640 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 28 | 0a12a65ee5557d1c | 3 | 2 | 4 | 15/16 | 1/8 | 0 | 3640 | 7 of 8 | 30 of 32 | float_trajectory_q15_attempt_cost |
| 29 | 32b36bf83cf94386 | 3 | 2 | 4 | 7/8 | 3/8 | 1/8 | 3645 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 30 | 036936ee657e3b5c | 3 | 2 | 4 | 7/8 | 3/8 | 3/16 | 3650 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 31 | a92dd66ca936d30e | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3660 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
| 32 | 2403f6837d3ac8fc | 3 | 2 | 4 | 15/16 | 3/8 | 0 | 3668 | 8 of 8 | 32 of 32 | float_trajectory_q15_attempt_cost |
float_trajectory_q15_attempt_cost
One table per job and one row per point, from its newest ledger line. A point counts as measured only under the code hash that measured it, so an older line from earlier code is left out; that rule and the retry policy are in the measurement ledger. Plan, job catalogue and invariants: docs/SIDETRACK-AUTOMATION.md in the harness repository. Designs these feed: docs/EPOCH2-DESIGN.md and docs/EPOCH3-DESIGN.md.
16 points
Closes: EPOCH2-DESIGN section 4: the dyadic PI gains rest on one measurement, and the classical alpha for a 3(2) pair is 1/3 rather than 1/4
Arithmetic: float64 only; no Q15 effects are included
| point | cycle | points | finished | max_error | rejection_rate | total_accepted | total_fevals | total_rejected |
|---|---|---|---|---|---|---|---|---|
| a1o2_b0o1 | 60 | 3 | 3 | 2.1067e-05 | 0.01309 | 377 | 1149 | 5 |
| a1o2_b1o16 | 60 | 3 | 3 | 1.94951e-05 | 0.01295 | 381 | 1161 | 5 |
| a1o2_b1o4 | 60 | 3 | 3 | 1.55843e-05 | 0.15748 | 428 | 1527 | 80 |
| a1o2_b1o8 | 60 | 3 | 3 | 1.82134e-05 | 0.01269 | 389 | 1185 | 5 |
| a1o4_b0o1 | 60 | 3 | 3 | 1.51356e-05 | 0.01663 | 414 | 1266 | 7 |
| a1o4_b1o16 | 60 | 3 | 3 | 1.23288e-05 | 0.01562 | 441 | 1347 | 7 |
| a1o4_b1o4 | 60 | 3 | 0 | n/a | n/a | n/a | n/a | n/a |
| a1o4_b1o8 | 60 | 3 | 3 | 8.24192e-06 | 0.01386 | 498 | 1518 | 7 |
| a1o8_b0o1 | 60 | 3 | 3 | 8.18309e-06 | 0.02534 | 500 | 1542 | 13 |
| a1o8_b1o16 | 60 | 3 | 3 | 2.7001e-06 | 0.01793 | 712 | 2178 | 13 |
| a1o8_b1o4 | 60 | 3 | 0 | n/a | n/a | n/a | n/a | n/a |
| a1o8_b1o8 | 60 | 3 | 0 | n/a | n/a | n/a | n/a | n/a |
| a3o8_b0o1 | 60 | 3 | 3 | 1.87781e-05 | 0.01018 | 389 | 1182 | 4 |
| a3o8_b1o16 | 60 | 3 | 3 | 1.69917e-05 | 0.00993 | 399 | 1212 | 4 |
| a3o8_b1o4 | 60 | 3 | 3 | 8.4292e-06 | 0.00797 | 498 | 1509 | 4 |
| a3o8_b1o8 | 60 | 3 | 3 | 1.51866e-05 | 0.00957 | 414 | 1257 | 4 |
8 points
Closes: EPOCH2-DESIGN section 2: the search-space note assumes a dyadic A admits both an order-p b and an order-(p-1) b_hat with a free parameter left, and that has not been counted
Arithmetic: exact over Fractions; no float and no Q15 arithmetic is involved
| point | cycle | b_hat_exact | b_hat_free | both_solvable | d_forced_zero | fsal | matrices | space_size | status |
|---|---|---|---|---|---|---|---|---|---|
| st3_s1 | 60 | 64 | 64 | 0 | 0 | 0 | 64 | 64 | ok |
| st3_s2 | 60 | 384 | 512 | 4 | 0 | 0 | 512 | 512 | ok |
| st3_s3 | 60 | 2048 | 4096 | 6 | 0 | 0 | 4096 | 4096 | ok |
| st3_s4 | 60 | 10240 | 32768 | 24 | 0 | 0 | 32768 | 32768 | ok |
| st4_s1 | 60 | 4096 | 4096 | 3646 | 0 | 0 | 4096 | 4096 | ok |
| st4_s2 | 60 | n/a | n/a | n/a | n/a | n/a | 0 | 262144 | capped |
| st4_s3 | 60 | n/a | n/a | n/a | n/a | n/a | 0 | 16777216 | capped |
| st4_s4 | 60 | n/a | n/a | n/a | n/a | n/a | 0 | 1073741824 | capped |
24 points
Closes: EPOCH2-DESIGN section 3: the roughly 2 LSB floor-bias figure for the Q15 error estimate is an estimate, and section 5 needs the scored tolerance ladder to sit above whatever that floor turns out to be
Arithmetic: Q15 fixed point throughout, on the pinned primitives in rk_harness.fixedpoint, with floor multiply; the reference float mode used to reproduce the frozen curve is float64
the bit scan is booked at its worst case and its branches cannot be priced by the pinned cost model, so the sequence figure is an upper bound rather than an exact count. The factor table is held in Q14 because a Q15 table cannot represent a factor of 1.0 at all; the encoding is an open item in EPOCH2-DESIGN section 4.
| point | cycle | points | finished | statuses | flattening_tol_lsb | mean_bias_lsb_pooled | mean_bias_lsb_total | overflow_rejected | tables_agree |
|---|---|---|---|---|---|---|---|---|---|
| battery_2rc_s1 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.00486615 | -2.03564 | 0 | false |
| battery_2rc_s2 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.00466504 | -1.89532 | 0 | false |
| battery_2rc_s4 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.0126933 | -1.943 | 0 | false |
| bicycle_lateral_s1 | 60 | 10 | 8 | attempt_cap, ok | 256 | -0.0168511 | -1.59451 | 0 | false |
| bicycle_lateral_s2 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.00646408 | -1.28954 | 0 | false |
| bicycle_lateral_s4 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.00734325 | -1.22897 | 0 | false |
| buck_converter_s1 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.0504386 | -1.8119 | 0 | false |
| buck_converter_s2 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.0101078 | -1.84062 | 0 | false |
| buck_converter_s4 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.016269 | -1.81396 | 0 | false |
| enzyme_qssa_s1 | 120 | 10 | 7 | attempt_cap, ok, step_underflow | 256 | -0.200438 | -1.8909 | 135 | false |
| enzyme_qssa_s2 | 120 | 10 | 7 | attempt_cap, ok, step_underflow | 512 | -0.12472 | -1.84144 | 136 | false |
| enzyme_qssa_s4 | 120 | 10 | 7 | attempt_cap, ok, step_underflow | 512 | -0.0850521 | -1.81784 | 140 | false |
| glucose_minimal_s1 | 60 | 10 | 7 | attempt_cap, ok, step_underflow | 512 | -0.0394501 | -1.91737 | 0 | false |
| glucose_minimal_s2 | 60 | 10 | 7 | attempt_cap, ok, step_underflow | 512 | -0.0237129 | -1.93891 | 0 | false |
| glucose_minimal_s4 | 120 | 10 | 7 | attempt_cap, ok, step_underflow | 512 | -0.0134941 | -1.90417 | 0 | false |
| pll_lock_s1 | 60 | 10 | 7 | attempt_cap, ok, overflow | 512 | -0.104358 | -1.94728 | 1 | false |
| pll_lock_s2 | 60 | 10 | 7 | attempt_cap, ok | 512 | -0.00373301 | -1.89591 | 0 | false |
| pll_lock_s4 | 60 | 10 | 8 | attempt_cap, ok | 512 | -0.00638674 | -1.82013 | 0 | false |
| robertson_scaled_s1 | 120 | 10 | 7 | attempt_cap, ok, step_underflow | 128 | -0.166833 | -1.88871 | 10 | false |
| robertson_scaled_s2 | 120 | 10 | 6 | attempt_cap, ok, overflow, step_underflow | 64 | -0.112156 | -1.88565 | 20 | false |
| robertson_scaled_s4 | 120 | 10 | 5 | attempt_cap, ok, overflow, step_underflow | 32 | -0.0732931 | -1.7575 | 32 | false |
| servo_load_step_s1 | 120 | 10 | 7 | ok, overflow, step_underflow | 256 | -1.58357 | -1.72787 | 3 | false |
| servo_load_step_s2 | 120 | 10 | 7 | ok, overflow, step_underflow | n/a | -1.52041 | -1.59191 | 3 | false |
| servo_load_step_s4 | 120 | 10 | 7 | ok, overflow, step_underflow | n/a | -1.45251 | -1.48171 | 4 | false |
8 points
Closes: EPOCH2-DESIGN section 5: the scored tolerance ladder is unchosen, and the published curve covers 3 of the 8 validation problems
Arithmetic: float64 only; no Q15 effects are included
| point | cycle | points | finished | statuses | fevals_max | fevals_min | rejection_rate_max |
|---|---|---|---|---|---|---|---|
| battery_2rc | 60 | 6 | 6 | ok | 421 | 28 | 0.0151515 |
| bicycle_lateral | 60 | 6 | 6 | ok | 1195 | 43 | 0.0434783 |
| buck_converter | 60 | 6 | 6 | ok | 3571 | 91 | 0.0166667 |
| enzyme_qssa | 60 | 6 | 6 | ok | 3301 | 1162 | 0.198905 |
| glucose_minimal | 60 | 6 | 6 | ok | 1675 | 49 | 0.0333333 |
| pll_lock | 60 | 6 | 6 | ok | 1603 | 52 | 0.0344828 |
| robertson_scaled | 60 | 6 | 6 | ok | 3883 | 2353 | 0.0425039 |
| servo_load_step | 60 | 6 | 6 | ok | 3145 | 3004 | 0.00667939 |
| document | state | what it holds |
|---|---|---|
| rk-work/validation/axes.json | not written yet | the cycles-to-tolerance rows for this class, and the measured floor of the Q15 error estimate that bounds how tight a tolerance can mean anything |
| rk-work/adaptive_archive/YYYY-MM-DD.jsonl | written | the lane's per-day search log: one line per candidate the lane enumerated, ranked against nothing |
| rk-work/schedule/shares.json | written | the per-cycle lane log summarised per method class |
The traceability rule lists key_findings.json, validation/results.json, benchmark/results.json, the side-track ledger with its artifacts, and the two lane elites documents, adaptive_archive/elites.json and implicit_archive/elites.json. Each document named above sits outside that list and says so in its own schema, so this page names it and reports whether it exists without publishing a number from it. Only the owner can add one to the list.