RESULTS.md

Model-Theory Normalization

RESULTS.md

Download original file

Original Markdown with line links

Retained results — v0.10

Cold Lean audit

PASS: nine modules; 159 named declarations; 105 theorems (78 generic, 27 examples). 17 expected-outcome tasks completed. Sum of subprocess runtimes: 52.769184 seconds. This is an execution receipt, not a comparative performance claim. All pinned accepted source hashes were unchanged. The positive control accepted and six invalid controls were rejected for their claims. Only standard axioms occurred.

Independent Python audit

PASS: 119,732 individual checks; 203 full discovery fixtures. All values use exact integer/Boolean arithmetic. Random wide cases use seed 2026092110. Native integer arithmetic supplies independent scalar reference results.

Check Count
unsigned_add 8,192
adder_cost_length 8,192
subtractor_borrow_identity 8,192
subtractor_cost_length 8,192
unsigned_multiply 4,096
multiply_quadratic_budget 4,096
unsigned_divmod 7,936
division_quadratic_budget 7,936
signed_add 4,225
signed_add_budget 4,225
signed_multiply_budget 4,225
signed_subtract_budget 4,225
signed_subtract 4,225
signed_multiply 4,225
signed_euclidean_divmod 8,224
signed_division_budget 8,224
residue_multiply 4,900
residue_wide_product_budget 4,900
residue_add 4,900
residue_add_budget 4,900
wide_padded_multiply 160
wide_padded_divide 160
wide_signed_divide 160
zero_divisor_rejected 3
malformed_word_rejected 4
frozen_parent_result_count 203
bit_backend_discovery_answer 203
bit_backend_canonical_basis 203
same_arithmetic_opcode_counts 203
same_arithmetic_total 203

The declared exhaustive bounds are: unsigned pairs 0..63, unsigned division dividends 0..255 and divisors 1..31, signed pairs -32..32, signed division dividends -128..128 and divisors 1..32, every residue pair at moduli 1..24. There are 160 separately seeded padded wide cases with widths drawn from 16,32,64,128,256. Finite tests do not replace generic Lean proofs.

Full discovery port results

All 203 ordered outputs and canonical bases equal the frozen originals. Every per-opcode arithmetic-event count matches, totaling 295,173 events. The new declared gate/cell total is 270,165,476. The maximum input magnitude width is 42 bits; padded output storage reaches 100 bits. These are different quantities.

Fixture Scalar events Gate/cell units Maximum operand width
scrambled_128bit.source 2,465 14,028,031 42
descending_torsion.source 696 1,045,360 25

The port deliberately re-encodes and decodes at scalar boundaries; conversion volume is recorded but its cost is not included. Extraction/Option interface charges are declared port conventions, not a whole-program Lean cost theorem. The 203 examples are inherited synthetic regressions, not new held-out discoveries.

Lineage and replay

The frozen parent archive SHA and all 1,395 listed payload hashes were verified. All 207 selected baseline payloads agree byte-for-byte with the parent. The recursive ZIP font scan found no font files. No parent Lean or full Mathlib rebuild is claimed. Final distribution and fresh-extraction results are in the separate EXTRACTED_REPLAY.json and packaging report, rather than inferred from this note.

Original file and source record

This copy matches the recorded source file. Its publication here does not change the work’s review status.

SHA-256: dd34ab5f44e1a2b3765bb9cf84454b4bead501831b853fa953bcab24e1105843

Original source on GitHub

Back to Model-Theory Normalization