Elamur Groupantsmath-bigbench-main
The groupant beside math/bigbench/src/main.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.
Facts
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | approximant |
| ast | partial | syntax.json |
| composition | partial | composition |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | partial bounded evidence | tests |
Obernants
The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial source derived model | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
failure_precedence | failure_precedence | partial source derived model | failure_precedence |
representation_transport | representation_transport | partial source derived model | representation_transport |
resource_boundary | resource_boundary | partial source derived model | resource_boundary |
domain_boundary | domain_boundary | partial source derived model | domain_boundary |
Model text
descriptant
Source-bound description
best(n,f) returns LAST result and MIN duration across n invocations; n=0 panics. Product tree multiplies1..n for n>=1, splits width>=16. big2048 uses32 wrapping LCG steps, shift right1, set low bit. main runs large factorial/power/GCD/modpow/harmonic comparisons.
Epoch 2026-10-31: prospective model before bounded synthetic tests; production behavior unchanged. Large benchmark main and large GCD/modpow unexecuted; no speed claim.
approximant
Forward relation
best(n,f) returns LAST result and MIN duration across n invocations; n=0 panics. Product tree multiplies1..n for n>=1, splits width>=16. big2048 uses32 wrapping LCG steps, shift right1, set low bit. main runs large factorial/power/GCD/modpow/harmonic comparisons.
Domain: Tests factorial1,2,7,17; harmonic3; pow3(8). fact(0) underflows and is outside domain; lo+hi overflow uncertified.
Executable Rust approximant unsupported: categorical/effect observations have no useful percentage metric yet. Future caps1/2/5/10% only; no automatic substitution.
invariant
Invariant and cost
Zero repetitions excludes best output. Main modpow comparison checks only last of100 outputs, not all bases.
Ordering: Elapsed collected before replacing output; cross-library assertions follow timed calls.
Cost: Product tree has O(n) factors with size-dependent multiplication; repeats multiply work. Conversion/GCD delegated; no timing performed this epoch.
composition
Composition and epoch memory
dashu/num library values compared via decimal representation, not type conversion.
Large benchmark main and large GCD/modpow unexecuted; no speed claim.
Partial syntax and finite tests do not establish all caller assumptions.
models
behavior_complement
Zero repetitions excludes best output. Main modpow comparison checks only last of100 outputs, not all bases.
reverse_reconstruction
Returned pair admits any n-output sequence ending with result and durations having given minimum, not necessarily timing of returned result. Decimal arithmetic outputs discard computation path.
failure_precedence
Elapsed collected before replacing output; cross-library assertions follow timed calls.
representation_transport
dashu/num library values compared via decimal representation, not type conversion.
resource_boundary
Product tree has O(n) factors with size-dependent multiplication; repeats multiply work. Conversion/GCD delegated; no timing performed this epoch.
domain_boundary
Tests factorial1,2,7,17; harmonic3; pow3(8). fact(0) underflows and is outside domain; lo+hi overflow uncertified.
tests
Evidence, 2026-10-31
Selected actual-source checks: cargo test -p bigbench wave28. Exact commands, numeric statuses, normalized outcomes, compiled assertion controls and restoration receipts are in the external wave28 mathcontrol run.
Large benchmark main and large GCD/modpow unexecuted; no speed claim.
No live teacher/API, trained model, corpus/held-out, GPU, DB or expensive benchmark run. Parent syntax generated twice identically. Scope remains partial.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/math/bigbench/src/main.rs.groupant/