RunMat
  • Pricing
RunMat
GitHub
GitHub
Sign inTry in browser
DesktopRuntimePhysicsCADAdministrationAgent

Get RunMat updates

New releases, guides and examples.

Product+
  • Runtime
  • GPU
  • Parallel
  • Compile
  • Interop and MEX
  • Simulation
  • MCAD
  • ECAD
  • Enterprise
  • AI integrations
  • Pricing
  • Download
Math+
  • Signal processing
  • Linear algebra
  • Statistics
  • Curve fitting
  • Differential equations
  • Control systems
  • Optimization
  • Finance
  • Data and tables
  • Plotting
Physics+
  • Simulation overview
  • Structural
  • Thermal
  • Electromagnetic
  • Joule heating
  • Thermal stress
  • Acoustics
  • Fluid flow
  • Conjugate heat transfer
  • Fluid–structure
  • What is FEA
Industries+
  • Aerospace
  • Automotive and transportation
  • Defense
  • Electronics and semiconductors
  • Energy
  • Industrial machinery
  • Life sciences
  • Quantitative finance
Learn+
  • Guides
  • Docs
  • Function reference
  • Language reference
  • Migration guide
  • Changelog
  • MATLAB Online
  • MATLAB AI agent
  • For students
Compare+
  • RunMat vs MATLAB
  • RunMat vs Ansys
  • RunMat vs SOLIDWORKS
  • RunMat vs Altium
  • RunMat vs Cadence
  • RunMat vs COMSOL
  • RunMat vs Python
  • RunMat vs Julia
  • RunMat vs Physics AI
  • RunMat vs Octave
  • Benchmarks
  • MATLAB alternatives
  • All comparisons
Company+
  • About
  • Blog
  • Contact
  • Source code↗
  • License
  • Privacy
  • Terms

Product

  • Runtime
  • GPU
  • Parallel
  • Compile
  • Interop and MEX
  • Simulation
  • MCAD
  • ECAD
  • Enterprise
  • AI integrations
  • Pricing
  • Download

Math

  • Signal processing
  • Linear algebra
  • Statistics
  • Curve fitting
  • Differential equations
  • Control systems
  • Optimization
  • Finance
  • Data and tables
  • Plotting

Physics

  • Simulation overview
  • Structural
  • Thermal
  • Electromagnetic
  • Joule heating
  • Thermal stress
  • Acoustics
  • Fluid flow
  • Conjugate heat transfer
  • Fluid–structure
  • What is FEA

Industries

  • Aerospace
  • Automotive and transportation
  • Defense
  • Electronics and semiconductors
  • Energy
  • Industrial machinery
  • Life sciences
  • Quantitative finance

Learn

  • Guides
  • Docs
  • Function reference
  • Language reference
  • Migration guide
  • Changelog
  • MATLAB Online
  • MATLAB AI agent
  • For students

Compare

  • RunMat vs MATLAB
  • RunMat vs Ansys
  • RunMat vs SOLIDWORKS
  • RunMat vs Altium
  • RunMat vs Cadence
  • RunMat vs COMSOL
  • RunMat vs Python
  • RunMat vs Julia
  • RunMat vs Physics AI
  • RunMat vs Octave
  • Benchmarks
  • MATLAB alternatives
  • All comparisons

Company

  • About
  • Blog
  • Contact
  • Source code↗
  • License
  • Privacy
  • Terms
GitHubXLinkedIn

© Dystr, Inc. · RunMat v0.7.1, Nov 17, 2026

RunMat® is a registered trademark of Dystr, Inc. MATLAB® is a registered trademark of The MathWorks, Inc. RunMat is not affiliated with, endorsed by, or sponsored by The MathWorks, Inc.

/
    • Solves and studies
    • Parallel and distributed execution
    • Evidence and artifacts
    • Results and trust
    • Simulation usage accounting
    • How each operation is counted
Docs>Physics>Run and results

Results and trust

A run result is useful when you can inspect it, explain it, reproduce it, and judge it against the workflow's standard.

RunMat records those signals in the result payload, saved artifacts, diagnostics, quality reasons, and provenance.

What A Run Returns

Every run returns common result data:

  • run_id,
  • a required run identity separating physics family, procedure, solution space, and exact formulation realization,
  • fields from the underlying FEA run payload,
  • optional domain payloads,
  • model_validity,
  • solver_convergence,
  • result_quality,
  • run_status,
  • publishable,
  • quality_reasons,
  • provenance.

Domain payloads can include modal frequencies, thermal temperature, gradient, heat-flux, energy-storage-rate, and axis-boundary summary data, transient snapshots, nonlinear convergence data, electromagnetic metrics, or coupled-family diagnostics.

Quality Gates

FieldMeaning
model_validityWhether model validation passed for the run.
solver_convergenceWhether the solver path converged under the selected policy.
result_qualityWhether domain-specific quality checks passed.
run_statusFinal status such as publishable, degraded, or rejected.
publishableBoolean accept/reject signal for callers.
quality_reasonsStable machine-readable reasons for warnings, degradation, or rejection.

Quality policy controls how strict the publishability decision is. A run can return useful diagnostics even when it is not publishable.

Diagnostics

Diagnostics explain observed behavior. They can describe solver fallback, modal residuals, nonlinear increment failures, electromagnetic source or boundary quality, thermal spread, CFD flow signals, and similar family-specific conditions.

Read diagnostics with quality reasons:

  • diagnostics say what was observed,
  • quality reasons say why that observation affected trust.

Same Study, Same Result

A study run is deterministic. For the same RunMat build on the same hardware class, the same checked study produces the same mesh digest, the same step run identities, saved solver states with identical content digests, and bit-identical assessed values on every run, whatever the local worker count or threads per worker. Worker placement, retries, queueing and measured stage times are recorded as execution evidence; they never enter the result identity. An execution policy changes where and how widely the work runs, not what it computes.

The guarantee is tested: crates/runmat-cli/tests/fea_determinism.rs runs one static study five times and with 1, 2 and 4 workers and 1 or 4 threads per worker, and compares the mesh digest, the saved state content digests and the assessed value.

Provenance

Provenance records how the run executed:

  • requested backend,
  • actual execution backend,
  • selected numerical backend,
  • precision mode,
  • deterministic mode,
  • typed solver method,
  • typed preconditioner,
  • typed fallback stage and reason.

Requested placement and actual execution are distinct. Uploading result fields to an accelerator does not turn a CPU solve into an accelerator solve. Any fallback is recorded with its typed stage and reason, and artifacts that violate their identity are rejected.

This matters for comparisons. Two runs can use the same model and operation but differ in backend, precision, fallback behavior, or solver policy.

Compare And Trend

Use persisted run ids to inspect later:

OperationPurpose
fea.results/v1Query one run with optional payload filtering.
fea.results_compare/v1Compare two persisted runs.
fea.trends/v1Summarize recent runs by family.

Comparisons and trends help turn individual solves into workflow decisions.

Trust Decision

Before treating a result as reliable, ask:

  1. Did the .fea file or model validate?
  2. Did the planned operation match the intended family?
  3. Did the selected backend and solver policy match expectations?
  4. Did the result pass quality gates?
  5. Are there quality reasons or diagnostics that affect use?
  6. Does the recorded validation evidence cover this use class?
  7. Is there V&V evidence for this model class, not just a successful run?
  8. Are artifacts available for review and reproduction?

Use Evidence & Artifacts for the record flow, Verification & Validation for correctness criteria, and Status for what the current release runs.

On this page
  • What A Run Returns
  • Quality Gates
  • Diagnostics
  • Same Study, Same Result
  • Provenance
  • Compare And Trend
  • Trust Decision