What this collection contains
The first campaign batch groups CT-01 battery parameter identification, an RLC equation problem and iron-rod heat-transfer problems. Its execution mode is published-reference comparison: MATLAB was not run.
The comparison table and report summarize retained source facts, local equations, results, gates and untested areas. The HPPC fitting evidence and interactive Studio forward example have different purposes; neither replays measured HPPC data.
Inspect and reproduce
Open the campaign report, comparison table and reference contract from this page. Follow an individual wiki example for its user workflow and limits.
From the repository root, node benchmarks/cross-tool/matlab-campaign/run.mjs emits a new result without replacing recorded evidence.
Use the campaign’s explicit --record workflow only when intentionally regenerating its evidence package. Retain source values, units, windows, gates, transformations and failures alongside the new results.
Provenance and release boundary
The campaign source notes record a September 30, 2026 documentation check. Live source URLs are mutable and no full source-page copy is retained; frozen contracts identify the imported facts.
The original campaign used an inventoried working tree with existing edits. The manifest hashes the Wasm binary and inputs; it does not establish that today’s Rust source built that binary. Its documentation requires an immutable rerun and review before external campaign release.
Published-reference agreement is not physical validation, general MATLAB equivalence or a performance advantage. Missing vendor traces and external execution evidence remain missing.
This is a historical result tied to its retained inputs, settings, reference version and source fingerprints. The current source tree or Wasm artifact may differ; a new Studio run creates new evidence and does not retroactively update this result.