bruno/homemaker-layout
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
homemaker-layout/examples/maple-court/coldstart-500000-s2.log

152 lines
9.4 KiB
Text
Raw Normal View History

coldstart health-centre seed 1 @ 500000: 9 fails (5h/4s) Cold-start re-baseline after the DESIGN.md 38.10/38.11 objective change. Single worker (avoids homemaker-py-b8g), scored by the shipped scorer from the programme directory. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MJ84Feep79Hhm3E4zZJmnB
2026-09-01 20:28:37 +01:00
WARNING: 338 m2 per storey needs ~70 m of daylit wall; the plot's non-private perimeter gives 55 m. Roughly 22 m2 of courtyard closes the gap. (DESIGN.md §38.3)
seed : /home/bruno/src/homemaker-layout/examples/maple-court/init.dom
programme : maple-court
budget : 500000
pop : 16
child_budget : 80
workers : 1
rng seed : 2
leaf sharing : True (factor=3)
superpose : False
multi_use : False
conn grade : False
use tiers : False
bridge circulation : False
ruin recreate : False
collapse in-search : True
shapecurve warmstart : False
shapecurve prune : False
output : /home/bruno/src/homemaker-layout/examples/maple-court/coldstart-500000-s2.dom
[ 80 evals] best 4.31031e-40 (fails 120) via construct/0
[ 1360 evals] best 5.82834e-38 (fails 118) via core_undivide noop
[ 1520 evals] best 1.66001e-37 (fails 117) via place_missing n -> rlrl
[ 1840 evals] best 3.62731e-36 (fails 107) via crossover rlll<->llrll
[ 2800 evals] best 3.14742e-35 (fails 104) via swap 0/llrr
[ 3600 evals] best 4.59088e-33 (fails 102) via place_missing tt -> rllr
[ 4160 evals] best 9.2155e-33 (fails 101) via swap 2/llrl
[ 4320 evals] best 9.28173e-33 (fails 101) via crossover lllll<->rlrll
[ 4400 evals] best 2.12991e-31 (fails 97) via level_retype 2/rlrl<->0/lrrll
[ 7440 evals] best 4.82473e-31 (fails 97) via core_divide rrrr (3 floors)
[ 8240 evals] best 8.3715e-31 (fails 95) via place_missing py -> rlrl
[ 10320 evals] best 1.73553e-29 (fails 90) via crossover rllll<->llrlr
[ 15120 evals] best 5.50371e-29 (fails 89) via swap 0/rrl
[ 16560 evals] best 1.36527e-28 (fails 87) via divide 1/rlrr
[ 17360 evals] best 1.3874e-28 (fails 87) via level_retype 0/lrrlr<->2/llrll
[ 18240 evals] best 4.52246e-26 (fails 80) via core_divide lrlr (3 floors)
[ 23200 evals] best 5.93841e-26 (fails 80) via divide 1/lll
[ 23360 evals] best 1.7257e-25 (fails 78) via place_missing py -> lrrr
[ 23760 evals] best 2.23563e-25 (fails 78) via crossover rlrrl<->lllll
[ 24240 evals] best 9.34508e-25 (fails 76) via crossover rlrrl<->rrlrr
[ 25920 evals] best 2.26323e-24 (fails 75) via crossover llllr<->rrrl
[ 27760 evals] best 2.14476e-24 (fails 74) via swap 0/rrll
[ 29520 evals] best 5.22955e-24 (fails 74) via level_fix ut1: lvl0/lrlrr → lvl2/rlll
[ 30320 evals] best 1.66122e-23 (fails 72) via place_missing n -> lrrr
[ 30560 evals] best 1.6949e-23 (fails 72) via rotate 0/rll
[ 31120 evals] best 1.92888e-23 (fails 72) via level_retype 0/lllr<->2/lrrr
[ 31840 evals] best 3.88073e-23 (fails 71) via rotate 2/llrl
[ 33520 evals] best 3.98572e-23 (fails 71) via undivide 0/rlrl
[ 35920 evals] best 7.53511e-23 (fails 70) via swap 0/llll
[ 36560 evals] best 7.94808e-23 (fails 70) via retype 1/rrlr->rc1
[ 37520 evals] best 8.1121e-23 (fails 70) via core_undivide noop
[ 37600 evals] best 1.56464e-22 (fails 69) via crossover rllrl<->rlrrr
[ 38160 evals] best 1.63217e-22 (fails 69) via divide 1/rllr
[ 39040 evals] best 1.64548e-22 (fails 69) via retype 1/rrrl->tt
[ 40400 evals] best 1.65094e-22 (fails 69) via retype 1/rrlr->O
[ 42400 evals] best 1.65488e-22 (fails 69) via level_retype 1/llll<->2/lrlr
[ 42560 evals] best 1.66645e-22 (fails 69) via rotate 0/rlr
[ 46320 evals] best 3.19675e-22 (fails 68) via undivide 2/llrl
[ 48240 evals] best 6.91742e-22 (fails 67) via crossover llll<->llrll
[ 48320 evals] best 6.95777e-22 (fails 67) via core_undivide noop
[ 48480 evals] best 7.14958e-22 (fails 67) via crossover rrlrr<->rlrrr
[ 48800 evals] best 7.16285e-22 (fails 67) via core_undivide noop
[ 50480 evals] best 7.17477e-22 (fails 67) via core_undivide noop
[ 50640 evals] best 1.39355e-21 (fails 66) via level_fix ph1: lvl2/rrlrlr → lvl0/lrr
[ 53200 evals] best 5.77265e-21 (fails 64) via rotate 0/rlr
[ 55600 evals] best 5.79924e-21 (fails 64) via crossover rlrll<->rlrrl
[ 58400 evals] best 5.81883e-21 (fails 64) via retype 1/rlrll->ur1
[ 62400 evals] best 9.76592e-21 (fails 63) via core_undivide noop
[ 63280 evals] best 1.07501e-20 (fails 63) via crossover rlrll<->rllrl
[ 65680 evals] best 1.15706e-20 (fails 63) via swap 0/rllr
[ 66320 evals] best 1.1586e-20 (fails 63) via level_retype 0/llll<->1/lrrl
[ 69680 evals] best 1.20599e-20 (fails 63) via core_undivide noop
[ 70800 evals] best 1.28103e-20 (fails 63) via retype 1/rlrlr->t
[ 71760 evals] best 1.82742e-20 (fails 62) via undivide 2/rllr
[ 73040 evals] best 2.56638e-20 (fails 62) via level_fix ur1: lvl1/rlrll → lvl0/lrr
[ 74000 evals] best 2.63265e-20 (fails 62) via swap 0/rrl
[ 74480 evals] best 6.24895e-20 (fails 61) via divide 2/rlll
[ 78960 evals] best 6.26967e-20 (fails 61) via swap 0/rlrl
[ 80560 evals] best 6.26967e-20 (fails 61) via core_undivide noop
[ 88560 evals] best 2.11893e-19 (fails 59) via crossover r<->r
[ 95840 evals] best 2.13379e-19 (fails 59) via crossover rrllr<->rrllr
[ 97360 evals] best 2.14849e-19 (fails 59) via core_undivide noop
[102080 evals] best 4.89088e-19 (fails 58) via swap 2/rllr
[115920 evals] best 5.35652e-19 (fails 58) via core_undivide noop
[116240 evals] best 5.39906e-19 (fails 58) via core_undivide noop
[117920 evals] best 5.40548e-19 (fails 58) via crossover llrr<->llrll
[125760 evals] best 8.66067e-19 (fails 57) via rotate 0/rlrl
[144640 evals] best 2.83792e-18 (fails 55) via crossover llrr<->lrl
[153040 evals] best 2.85323e-18 (fails 55) via core_undivide noop
[155600 evals] best 2.86486e-18 (fails 55) via core_undivide noop
[157760 evals] best 3.22807e-18 (fails 55) via divide 0/llrlr
[158400 evals] best 5.72902e-18 (fails 54) via undivide 2/rrlrl
[163600 evals] best 7.12352e-18 (fails 54) via swap 2/rllr
[164000 evals] best 7.23944e-18 (fails 54) via rotate 0/rlrr
[165680 evals] best 7.27269e-18 (fails 54) via level_retype 0/lrll<->2/llrl
[167840 evals] best 7.30749e-18 (fails 54) via swap 0/rrr
[168720 evals] best 7.3439e-18 (fails 54) via retype 0/rlrlr->ef1
[188800 evals] best 1.43357e-17 (fails 53) via swap 0/llrl
[193840 evals] best 3.11483e-17 (fails 52) via rotate 0/lr
[194080 evals] best 3.13028e-17 (fails 52) via level_retype 0/lrr<->2/rllrl
[196800 evals] best 3.13837e-17 (fails 52) via core_undivide noop
[207920 evals] best 3.1393e-17 (fails 52) via core_undivide noop
[208160 evals] best 3.17987e-17 (fails 52) via crossover rrll<->rrll
[222320 evals] best 3.72977e-17 (fails 52) via core_divide lrlr (3 floors)
[223920 evals] best 6.33334e-17 (fails 51) via swap 0/llrl
[227280 evals] best 7.07114e-17 (fails 51) via level_fix k1: lvl1/lrlrr → lvl0/lrr
coldstart maple-court seed 1 @ 500000: 73 fails (17h/56s) Cold-start re-baseline after the DESIGN.md 38.10/38.11 objective change. Single worker (avoids homemaker-py-b8g), scored by the shipped scorer from the programme directory. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MJ84Feep79Hhm3E4zZJmnB
2026-09-02 10:52:55 +01:00
[233200 evals] best 7.61239e-17 (fails 51) via retype 2/rrlrr->O
[239680 evals] best 1.33703e-16 (fails 50) via level_fix rc1: lvl2/lrlrr → lvl0/lrr
[255520 evals] best 1.34474e-16 (fails 50) via level_fix rc1: lvl2/lrlrr → lvl0/lrr
[255920 evals] best 1.34475e-16 (fails 50) via retype 1/lllr->ph1
[267200 evals] best 1.43929e-16 (fails 50) via level_fix k1: lvl1/lrlrr → lvl0/lrr
[280880 evals] best 1.44209e-16 (fails 50) via swap 0/rlrl
[285760 evals] best 1.44897e-16 (fails 50) via crossover rlr<->rlr
[290640 evals] best 1.46495e-16 (fails 50) via rotate 0/rlrr
[294800 evals] best 1.4687e-16 (fails 50) via swap 2/rrlr
[302240 evals] best 1.47593e-16 (fails 50) via core_undivide noop
[303200 evals] best 1.51095e-16 (fails 50) via crossover rrrl<->llrlr
[308400 evals] best 1.51414e-16 (fails 50) via crossover lrll<->rrlrr
[310560 evals] best 1.51745e-16 (fails 50) via retype 1/rrlr->m
[313120 evals] best 1.52084e-16 (fails 50) via retype 1/rrll->ur2
[314960 evals] best 1.52441e-16 (fails 50) via level_retype 1/rrll<->0/rlrrr
[316240 evals] best 1.5514e-16 (fails 50) via core_undivide noop
[316400 evals] best 2.98902e-16 (fails 49) via core_undivide noop
[319440 evals] best 3.04842e-16 (fails 49) via level_retype 0/rrllll<->1/lrrrl
[320800 evals] best 3.06083e-16 (fails 49) via swap 0/rlrl
[323040 evals] best 3.07396e-16 (fails 49) via core_undivide noop
[324320 evals] best 3.17316e-16 (fails 49) via undivide 0/rrrl
[326800 evals] best 3.18549e-16 (fails 49) via crossover rlrll<->lrll
[327600 evals] best 3.19798e-16 (fails 49) via crossover rrllll<->rlllrr
[333040 evals] best 3.22121e-16 (fails 49) via core_undivide noop
[334400 evals] best 3.29134e-16 (fails 49) via crossover rrrr<->lrlrl
[337200 evals] best 3.29777e-16 (fails 49) via level_retype 1/rrrr<->0/rllll
[349760 evals] best 3.30314e-16 (fails 49) via core_undivide noop
[353440 evals] best 3.34503e-16 (fails 49) via undivide 0/rrrl
[353680 evals] best 5.9728e-16 (fails 48) via rotate 0/rrl
[369600 evals] best 5.98058e-16 (fails 48) via core_undivide noop
[371920 evals] best 5.98874e-16 (fails 48) via retype 0/rllr->gh1
[377520 evals] best 5.99301e-16 (fails 48) via core_undivide noop
[381760 evals] best 5.99551e-16 (fails 48) via crossover rllr<->rllr
[390160 evals] best 6.19047e-16 (fails 48) via crossover rlrl<->rlrr
[403440 evals] best 6.54129e-16 (fails 48) via swap 2/rrlr
coldstart harbor-house seed 2 @ 500000: 42 fails (12h/30s) Cold-start re-baseline after the DESIGN.md 38.10/38.11 objective change. Single worker (avoids homemaker-py-b8g), scored by the shipped scorer from the programme directory. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MJ84Feep79Hhm3E4zZJmnB
2026-09-02 17:53:41 +01:00
[408960 evals] best 6.55008e-16 (fails 48) via retype 0/rlrll->n
[413440 evals] best 6.55469e-16 (fails 48) via crossover rrllll<->rrllr
[413680 evals] best 6.55739e-16 (fails 48) via core_undivide noop
[445280 evals] best 6.56519e-16 (fails 48) via core_undivide noop
[446960 evals] best 6.57378e-16 (fails 48) via retype 2/llrl->k1
[447840 evals] best 6.57828e-16 (fails 48) via retype 0/rlrll->ut1
[451040 evals] best 6.58047e-16 (fails 48) via core_undivide noop
[471120 evals] best 6.81318e-16 (fails 48) via rotate 0/rlrl
[471840 evals] best 6.81796e-16 (fails 48) via crossover rrlrl<->rrlrl
[472960 evals] best 6.83308e-16 (fails 48) via swap 0/rlllr
[488880 evals] best 2.62087e-15 (fails 46) via retype 2/rlllrr->C
Reference in a new issue Copy permalink