{"status": "partial: gas diet + 16-unit zero-output widening built, compiled and proven identical to HA on regression seeds; NO training, NO replica/port gates (A/B/C) and NO fresh-seed real confirmation done", "hashes": {"HA_code_sha256": "99e3a2eceae0e5543d27473859e10b005d0ce0aa7067f2e1a24d1e0e4a7390fe", "T_json_sha256": "ebd9de9bf6af6e55d5e308fd15dbb88edec11fd684891d02a30aec185acd19b3 (T artifact code sha256, verified)", "diet8_sha256": "383b9857ec752fb1256e13b914dcba9170c99d314d9335437b2f8c958e0dd07e", "diet16_sha256": "3947c67a7b8c4548b4184652505d15a7a5a1dbd4597e8257abbcbbd79de6bee5", "HA_artifact_id": "2801efc8190e86318dd922d4a7179f63684d77a226047fc2f3c78232963eb690"}, "diet_code_zb64": "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", "diet_code_sha256": "3947c67a7b8c4548b4184652505d15a7a5a1dbd4597e8257abbcbbd79de6bee5", "diet8_code_zb64": "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", "diet8_code_sha256": "383b9857ec752fb1256e13b914dcba9170c99d314d9335437b2f8c958e0dd07e", "params": {"hidden_units": 16, "W1": [[-0.042655, 1.389981, -0.435022, 2.131939, -2.307678, -0.134711, -1.563002, 0.838218, -3.867643, -0.287527, -0.014871, 0.138925, 0.547082], [0.840878, 1.015656, 1.905167, 4.344855, 1.49152, 0.813268, 0.140392, -0.166508, -0.344279, 0.280778, -0.262185, -0.623571, 0.543047], [2.339559, 1.927192, -1.149669, 0.419394, 2.422033, 2.797237, -0.184718, 2.200018, 2.191446, -1.193034, -0.027589, -1.571561, -1.126364], [-1.436015, -0.245775, 0.519839, 0.840096, -1.06768, 0.941901, -0.447013, 1.415013, -1.647625, 0.825277, 0.681432, -1.416949, -1.754078], [1.014618, -1.515656, 3.692017, -0.716903, -2.472535, -0.750627, -1.234693, -1.457361, 0.437293, 0.064478, -0.154363, -3.496125, 0.238174], [1.237463, -0.770777, 0.793692, -1.766367, 2.499688, 1.239927, 0.842836, -0.357134, -2.299169, 0.131642, 0.337489, 3.91456, 0.554438], [-2.559287, 0.779086, -0.480982, -2.426775, 2.235989, 0.776522, -0.2552, -0.074918, 1.361255, -0.489339, 0.456458, 1.516253, -0.226631], [-3.164297, -1.575901, 2.530196, 3.744943, -1.077146, -1.736719, -0.508633, 1.762971, 0.752575, -0.725826, 2.462294, -2.054677, 1.028772], [-0.042655, 1.389981, -0.435022, 2.131939, -2.307678, -0.134711, -1.563002, 0.838218, -3.867643, -0.287527, -0.014871, 0.138925, 0.547082], [0.840878, 1.015656, 1.905167, 4.344855, 1.49152, 0.813268, 0.140392, -0.166508, -0.344279, 0.280778, -0.262185, -0.623571, 0.543047], [2.339559, 1.927192, -1.149669, 0.419394, 2.422033, 2.797237, -0.184718, 2.200018, 2.191446, -1.193034, -0.027589, -1.571561, -1.126364], [-1.436015, -0.245775, 0.519839, 0.840096, -1.06768, 0.941901, -0.447013, 1.415013, -1.647625, 0.825277, 0.681432, -1.416949, -1.754078], [1.014618, -1.515656, 3.692017, -0.716903, -2.472535, -0.750627, -1.234693, -1.457361, 0.437293, 0.064478, -0.154363, -3.496125, 0.238174], [1.237463, -0.770777, 0.793692, -1.766367, 2.499688, 1.239927, 0.842836, -0.357134, -2.299169, 0.131642, 0.337489, 3.91456, 0.554438], [-2.559287, 0.779086, -0.480982, -2.426775, 2.235989, 0.776522, -0.2552, -0.074918, 1.361255, -0.489339, 0.456458, 1.516253, -0.226631], [-3.164297, -1.575901, 2.530196, 3.744943, -1.077146, -1.736719, -0.508633, 1.762971, 0.752575, -0.725826, 2.462294, -2.054677, 1.028772]], "b1": [1.446428, -2.846664, -3.039159, -0.668153, -0.67968, -1.68127, -2.09327, -0.152766, 1.446428, -2.846664, -3.039159, -0.668153, -0.67968, -1.68127, -2.09327, -0.152766], "W2": [[0.532461, -1.789106, -0.079941, -0.398251, -3.782992, -0.373606, -1.339628, 0.932635, 0, 0, 0, 0, 0, 0, 0, 0], [0.237721, -1.524133, -0.899179, 0.243723, 0.796095, -0.493254, -0.436366, 0.913103, 0, 0, 0, 0, 0, 0, 0, 0]], "b2": [0.654604, 1.544139], "weight_fixed_point_scale": 1000000, "feature_fixed_point_scale": 1000000, "output_unit_fee": 0.0001, "feature_order": ["sigmaHat", "lambdaHat", "sizeHat", "toxEma", "actEma", "direction", "(spot-staleHat)/staleHat", "signed_trade_ratio", "arb_probability", "first_in_step", "uncertainty", "min(silent,8)/8", "anchor_mean"], "feature_multipliers": [1000, 2, 500, 500, 500, 1, 500, 500, 2, 2, 1000000, 1, 500], "feature_offsets": [-1, -1, -1, -1, -1, -1, 0, 0, -1, -1, -1, 0, -1], "feature_clip": [-4, 4], "hidden_relu_clip": [0, 4], "toxicity_anchor": {"count_cap": 674, "linear_k_wad": 712696014807650816, "quadratic_k_wad": 1295240753826310656}, "cma": {"seed": 21, "population": 16, "dimension": 130, "initial_sigma": 0.2, "W1_initial_rng": "numpy.default_rng(21).normal(0,0.35,104)", "other_initial_values": 0, "generations": 413, "train_seeds": [22000000, 22105727], "train_markets_per_generation": 256, "monitor_seeds": [30000000, 30002047], "monitor_every_generations": 10, "selection": "best historical training-block proposal rescored on monitor; retain maximum monitor gain"}}, "params_note": "16 hidden units: units 8..15 copy HA units 0..7 (W1,b1); new W2 columns and b2 unchanged (new output weights 0). Fixed-point scales as HA (weights 1e6). A zero-weight literal is constant-folded by solc, so retraining must edit literals in _head; gas of live weights measured separately (D16p).", "design": {"storage": "17 storage slots packed into 6: slot0 bid|ask|ts|stepCount; slot1 dir|act|sigma|lambda (64b each); slot2 size|tox|uncertainty; slot3 pHat|fastHat (128b); slot4 prevSpot(128)|runMean(64)|sampleCount(16)|d1(2); slot5 flow (unchanged 3x80). Bounds: all values < 2^64 by caps (MAX_FEE 1e17, 2*WAD, LAMBDA_CAP 1.82e18, RET_CAP, TOX_CAP, 1e14 uncertainty); prices in 128 bits. No truncation masks added (would cost gas); relied on bounds + equality tests.", "pow": "_powWad(decay,e) for e=1..8 replaced by exact lookup tables generated from the same wmul chain (DIR .75, ACT=SIZE .53, TOX .91); retention 0.999^silent keeps the loop minus last squaring, fast path e==1; density pow4 = two wmul; silence weight with g=1 is the constant 0.2e18.", "sqrt": "_isqrt: same bit-length seed, 6 Newton steps for x<2^108 (all callers <~2^107), 7 above, then min(r,x/r); saves one step only. Gas gains are almost entirely from storage packing."}, "speed": {"note": "Measured: run_batch gives ~11.5-15.6 s per 10000-step market with no multi-process speedup on this machine (user~=real; 8 processes gave 32 full markets in ~8 min). Replica/port not built, so no replica markets/s."}, "restart": {"weights": "params above (HA weights, 16 units, zero new outputs)", "ES_seed": "not run", "settings": "no training performed"}, "curve": "none: no training", "gates": {"A": "not done", "B": "not done", "C": "not done", "note": "replica/port gates and search were skipped for lack of time; the only runs were EVM regression and gas."}, "seeds": {"identity_full_length_10000_steps": "seeds 0..15, HA vs diet8: pnl, edge, normalizer pnl/edge, arb/retail volume, average fees and sha256 of all 10000 per-step submission fees identical (0 mismatches)", "identity_regression_500_steps": "seeds 0..99 (configs per MatchRunner._build_configs with seed=i, n_steps overridden to 500, run_batch vs VanillaStrategy): diet8 and diet16 each 0/100 mismatches vs HA in the same fields incl. per-step fee hash. DEVIATION: 500 steps not 10000, because markets run ~12-15 s each on a single effective core.", "gas_smoke": "block tail seeds 7132760..7132767, n_steps overridden to 2000 (DEVIATION: not full length, not the 64-seed tail): HA, diet8, diet16, diet16-live-weights probe"}, "replica_validation": "not done", "real_paired": {"vs_T": "not run", "vs_HA": "not run; identity vs HA shown instead (every compared field equal, so mean_diff=0, se=0, worst5=[] on the compared seeds)", "note": "no fresh-seed 512 confirmation vs 30bps vanilla was possible in the time/CPU available"}, "retail_arb_deltas": "0 vs HA on compared seeds (volumes identical); T/HA vs T not measured", "gas": {"full_length_seeds_0_15": {"HA": {"afterSwap_max": 207864, "afterSwap_mean": 131305, "afterInit_max": 243147, "oog": 0}, "diet8": {"afterSwap_max": 112130, "afterSwap_mean": 88074, "afterInit_max": 132588, "oog": 0}}, "regression_500_steps_seeds_0_99": {"HA": {"afterSwap_max": 233608, "mean": 131657, "afterInit": 243147, "oog": 0}, "diet8_and_16": {"afterSwap_max": 111960, "mean": 88246, "afterInit": 132588, "oog": 0}}, "tail_2000_steps_seeds_7132760_7132767": {"HA": {"max": 207886, "mean": 130997}, "diet8_and_16": {"max": 111910, "mean": 87845, "afterInit": 132588}, "diet16_live_nonzero_weights_probe": {"max": 113219, "mean": 89211, "afterInit": 132688}, "oog_all": 0, "other_failures": 0}, "headroom": "16 live units cost only ~+1.3k gas/call; afterSwap peak ~113k vs limit 245k, so far more than 16 units (roughly 100+ more 13-input units at ~160 gas each) would fit. Reference HA peak over 512 real seeds was 231,381; diet peak over that many seeds not measured (peak in our samples fell 46%).", "source": "amm_sim_rs.gas_stats() per process, Submission role"}, "commands": ["sh scratch/setup.sh (goal setup verbatim, passed incl. amm-match validate)", ". scratch/env.sh && python -I scratch/w/gen.py (generator: HA.sol -> D8/D16/D16p)", "solc 0.8.24 viaIR optimizer 200 via SolidityCompiler.compile for each variant", "NSTEPS=500 python -I scratch/w/par.py HA,D8,D16 0 100 10", "NSTEPS=2000 python -I scratch/w/par.py HA,D8,D16,D16p 7132760 7132768 2"], "notes": ["Inputs verified: HA artifact file sha256 equals URL id; HA code sha256 99e3a2ec... matches; T code sha256 matches. Replica and RS2 artifacts downloaded but not used.", "Not delivered: port_zb64 (no port built), training, replica/real confirmation vs T. Whole workspace was run inside test/scratch (not committed).", "The diet16 deliverable has behavior identical to HA by construction (zero output weights); it is a gas-lean restart point for training, not an improvement.", "Compile failed first (stack too deep) with an unpacked-locals version; fixed by scoping locals and packing d1/runMean/count into one local."]}