# Validation

Channel: release; package: 1.8.0; source: 7d1e97331656dd74ad0f5471653ac0c5effdff70

Page: https://remilab.ai/rfx/versions/v1.8.0/validation/
Source: https://github.com/bk-squared/rfx/blob/7d1e97331656dd74ad0f5471653ac0c5effdff70/docs/public/validation/index.mdx

Use this page to select a documented configuration and the evidence required for
an rfx result. Begin with the
[Recommended Configuration](https://remilab.ai/rfx/versions/v1.8.0/validation/reference-lane/) and then check
[Support Boundaries](https://remilab.ai/rfx/versions/v1.8.0/api/support-boundaries/) for the exact mesh, source,
port, calculator, and observable restrictions.

## Configuration and required validation

| Configuration | Required validation |
|---|---|
| uniform Cartesian Yee RF/FDTD workflows | use the documented boundaries, sources, observables, convergence checks, and analytic or independent reference |
| rectangular waveguide S-matrices | use `compute_waveguide_s_matrix(...)` only for the documented guide geometry, modes, frequency range, normalization, reference planes, and placement |
| lumped, wire, microstrip-line, and coaxial-line calculations | use the calculator for that port family and remain within its stated limits |
| differentiable design loops | treat the optimization quantity as a proxy and validate the final RF observable separately |
| other repository code | treat RF accuracy as unvalidated unless a public guide and support entry define the configuration and evidence |

## Evidence, artifacts, and reports

Plots, Touchstone files, HDF5 snapshots, CSV tables, and report bundles are
useful review artifacts. They become validation evidence only when paired with
a documented analytic result, dump replay, external-solver comparison,
convergence study, or benchmark acceptance criterion.

The measured cases themselves — every script under `validation/crossval/` with
its reference, gate and reproduce command — are listed in
[Benchmarks](https://remilab.ai/rfx/versions/v1.8.0/guide/benchmarks/).