meshing.evidence — Return the validated stage, resource, and identity evidence of a solver mesh.
evidence = meshing.evidence(mesh) returns a meshing.Evidence object describing how a meshing.Mesh was generated and what each meshing stage proved.
Syntax
evidence = meshing.evidence(mesh)Inputs
mesh- Type
Any- Required
- Yes
- Default
- —
- Description
- meshing.Mesh from meshing.generate or meshing.load.
Returns
evidence- Type
Any- Description
- meshing.Evidence object with stage, resource, and identity records.
Errors
RunMat:meshing:InvalidInput- When
- A meshing object, path, option, artifact, or evidence contract is invalid, or meshing cannot produce a result that satisfies the request.
- Message
meshing operation failed
Behavior
Identity properties: artifact_digest is the canonical digest of the mesh the evidence belongs to and equals mesh.canonical_digest; deterministic_seed is the seed used; geometry_revision is the revision of the source geometry. Totals: stage_count, generated_nodes, generated_elements, and wall_time_ms. All counters are exact uint64 values.
resources is a struct of aggregate measurements: generated_nodes, generated_elements, peak_memory_bytes, peak_scratch_bytes, wall_time_ms, artifact_bytes, search_work, maximum_recursion_depth, and iterations.
stages is a 1-by-S cell of structs in execution order. Each has stage (for example "curve_mesh"), progress counters completed_work and estimated_work, peak_memory_bytes, peak_scratch_bytes, search_work, iterations, elapsed_time_ms, cancellation_checkpoints, and invariants, a cell of structs with id, passed, and, when the check is quantitative, measured, required, and unit.
The evidence is validated against the mesh whenever a meshing.Mesh is created or loaded, so an admitted mesh never carries a failed invariant. Timing and memory values describe the run that produced the mesh and can differ between otherwise identical runs; the mesh itself does not. When the native host reuses accepted stage results from an earlier run, those stages keep the counters and timings measured when they were produced.
Use in a study
Record the evidence with analysis results to show which request, seed, and geometry revision produced the mesh and which checks it passed. For Tet10 meshes, the evidence is the authority for curved-element Jacobians and achieved boundary accuracy; meshing.quality reports straight-sided corner measures only.
Errors
Every failure reports RunMat:meshing:InvalidInput, with a message that names the operation and the specific cause.
The input must be a meshing.Mesh.
Example
Confirm what the evidence proves about a mesh
part = geometry.load('box.brep', 'Units', 'm');
solverMesh = meshing.generate(part, 'TargetSize', 0.5, 'MaximumDeviation', 0.05, 'ElementOrder', 'tet4', 'Seed', 42);
ev = meshing.evidence(solverMesh);
stageNames = cellfun(@(s) s.stage, ev.stages, 'UniformOutput', false)
allPassed = all(cellfun(@(s) all(cellfun(@(i) i.passed, s.invariants)), ev.stages))Using meshing.evidence with coding agents
Open a RunMat example with live inputs, then ask the agent to explain how meshing.evidence changes the result.
Run a small meshing.evidence example, explain the result, then change one input and compare the output.
FAQ
Why does wall_time_ms change between runs with the same seed?⌄
Resource measurements describe the execution, not the result. The mesh and its canonical digest are reproducible; timing and memory are not.
Related Meshing functions
meshing.boundaryFaces · meshing.elements · meshing.generate · meshing.load · meshing.nodes · meshing.quality · meshing.regions · meshing.save
Related guides
Source you can read
Every RunMat function's source is public. Read exactly how meshing.evidence is executed, line by line, in Rust.
- View the source for meshing.evidence in Rust on GitHub
- Learn how the RunMat runtime works
- Found a bug? Open an issue with a minimal reproduction.
About RunMat
RunMat is a runtime that executes MATLAB-syntax code fast on any GPU. The runtime is free for individuals, including freelancers; students and faculty at an academic institution, for teaching or research there; contributors to an open-source project, for work on that project; organizations with 250 or fewer people and US$10M or less in annual revenue; and anyone, for non-production use. See runmat.com/license.
- RunMat automatically optimizes your math for GPU execution on Apple, Nvidia, and AMD hardware. No code changes needed. Simulations that took hours now take minutes.
- Start running code in seconds. RunMat runs in the browser, on the desktop, or from the CLI. No license server, no IT ticket.