Elamur Groupantsmlab-numerics-numeric-ode
The groupant beside mlab/src/builtins/numeric.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 |
| cost | symbolic source derived | invariant |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | executed bounded checks | 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 |
arity_failure | arity_failure | partial source derived model | arity_failure |
rhs_shape | rhs_shape | partial source derived model | rhs_shape |
step_acceptance | step_acceptance | partial source derived model | step_acceptance |
zero_interval | zero_interval | partial source derived model | zero_interval |
Model text
descriptant
mlab-numerics-numeric-ode
ode45 validates output arity/options, gets real tspan/y0 storage, evaluates initial RHS, then adapts Dormand-Prince step using scaled embedded error. Accepted steps append all times when two tspan points, otherwise only requested output targets; RHS must return real vector with neq cells.
Modeling epoch 2026-10-31
Prospective source reading preceded inline tests. This node records implementation behavior, not a complete MATLAB compatibility contract. Models retain limits discovered in source reading, including real-only projections and warning-only result gates. Synthetic CPU fixtures are independent of frozen suites. No external effects, teacher, corpora or benchmark main are executed.
approximant
Forward relation
ode45 validates output arity/options, gets real tspan/y0 storage, evaluates initial RHS, then adapts Dormand-Prince step using scaled embedded error. Accepted steps append all times when two tspan points, otherwise only requested output targets; RHS must return real vector with neq cells.
Executable Rust surrogate gap: no calibrated replacement is attached. Exact test calls are evidence for the reference, not a fast approximant. Numerical approximants would need a named domain, zero convention and metric at supported caps 1/2/5/10%; categorical errors/effects have no percent surrogate here.
invariant
Invariants and cost
t is directed toward final endpoint on admitted monotone tspan; stored state matrix uses rows=time, columns=equations. No monotonicity validation for arbitrary tspan.
Cost
At most1e6 attempted steps; each executes stages plus extra final RHS, O(neq) stage work and O(neq*outputs) storage; callback work external.
Allocation failure, unchecked usize dimension arithmetic and IEEE nonfinite propagation are outside finite fixture guarantees. No measured throughput or global accuracy claim.
models
Typed source models: mlab-numerics-numeric-ode
behavior_complement
nargout1 excludes successful struct output. RHS length mismatch excludes stepping. Rejected steps exclude trajectory append but not callback effects. Nonfinite error or too-small step errors; no dense-output interpolation.
reverse_reconstruction
A returned table fixes sampled states at times; compatible RHS values must produce accepted stage updates within code's local estimates, but off-stage fields remain unconstrained. Equal-time output can come from any RHS passing initial shape validation.
arity_failure
Empty args with nargout1 fails struct-output policy before missing-function argument.
rhs_shape
eval_rhs accepts any real Mat with neq cells, then interprets vector order; complex storage rejects.
step_acceptance
err<=1 appends accepted y; next factor is5 if zero else clamp(.9*err^-0.2,.2,5).
zero_interval
Initial RHS still evaluates when tf==t0; while loop skips and returns initial row.
tests
Evidence and limits
Selected wave30_tests exercise this source; fragment-specific limits are stated in its operators.
Native command specifications, numeric exits, profile-paired logs, compiled controls and exact-restoration receipts are retained externally in the wave30 numerics run. Tests exercise actual source, not a copied model. The selected finite fixtures do not prove full convergence, backend accuracy on all matrices, resource exhaustion or side-effectful callback behavior.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/mlab/src/builtins/numeric.rs.groupant/numeric-ode.groupant/