Qualified bounded
Accepted only for a recorded combination, model and observable scope.
Verification & validation
Follow each claim to its evidence.
Open a recorded case, inspect its model and results, then follow the versioned contract, raw record and decision. A passing calculation supports only its declared scope.
Recorded cases & original sources
Inspect the model. Then the result.
These eight case pages connect assumptions, formulation, observables, thresholds, source artifacts and final scope. Published values summarize existing records; they are not new solver executions.
4 recorded cases
External numerical correlation
Four frozen one-element linear static cases against Code_Aster 18.1.0. Agreement is limited to comparable observables on the same mesh.
Owner accepted with limitations · WP12 R2TET4 · Code_Aster correlation
One homogeneous linear static element, compared on the same mesh with Code_Aster 18.1.0. Historical development cycle evidence retained in the published source.
→Owner accepted with limitations · WP12 R2HEX8 · Code_Aster correlationOne homogeneous linear static element, compared on the same mesh with Code_Aster 18.1.0. Historical development cycle evidence retained in the published source.
→Owner accepted with limitations · WP12 R2TET10 · Code_Aster correlationOne homogeneous linear static element, compared on the same mesh with Code_Aster 18.1.0. Historical development cycle evidence retained in the published source.
→Owner accepted with limitations · WP12 R2HEX20 · Code_Aster correlationOne homogeneous linear static element, compared on the same mesh with Code_Aster 18.1.0. Historical development cycle evidence retained in the published source.
→1 recorded case
Gyroscopic and Campbell evidence
Formulation and tracking checks, an analytical oscillator and internal BEAM2 refinement, with ambiguity preserved.
3 recorded cases
Historical performance records
Named structured TET4 workloads and environments. Each record retains its execution source and limitations.
Historical environment specific observationsPETSc · historical 1M and 3M workloads
Structured TET4 static workloads recorded with named PETSc and MPI configurations. Different iterations and configurations are separate observations.
→Historical bounded Silver evidence · two complete runsPETSc · 5M Silver replaysExact structured TET4 linear static workload on the frozen single host configuration. Execution source: 04b07e00ddfe0b339b5790493a607ec902b1ed80.
→Historical bounded solve observationPETSc · bounded 10M solve observationA completed C3 structured TET4 linear static solve. Its audit and replay records must be read separately; a completed solve does not establish every qualification level.
→5M run 19.8472e−11free relative residual
5M run 11.3908e−9equilibrium relative error
5M run 11.5437e−14energy relative error
5M Silver2 / 2complete runs
Historical structured TET4 observations use named configurations. They establish no general scaling, mixed-mesh, nonlinear or hardware-independent capacity. Values are rounded from the frozen Silver replay record ↗.
Evidence method
From a model to a bounded decision.
Public records preserve the original execution identity. Viewing a result in a newer site release does not mean that campaign was rerun.
- 01Define
Contract
Declare the model, observables, thresholds and failure policy before execution.
- 02Record
Execution
Keep the exact source, inputs, environment and raw results identifiable.
- 03Challenge
Independent checks
Distinguish analytical checks, external numerical correlation and physical measurements.
- 04Reproduce
Replay
Compare recorded states, digests and observables without changing the frozen gates.
- 05Conclude
Decision
Read the final decision, accepted scope and unresolved limitations separately.
Experimental or bounded evidence
Implementation and selected results are available. A successful check does not automatically promote maturity.
Not validated or research only
The evidence needed for the proposed scope is not established. Historical failures remain visible.
Package 0.2.11 retains registry 0.2.8 and separate route decisions. Read the authoritative capability definitions ↗
Campaign notes
Read decisions with their source cycle.
Implementation history, execution evidence, review and release status are separate records.
GYRO 01–06 cover analytical, formulation, tracking and internal mesh checks. A high-frequency pair remains ambiguous at 100 rad/s.
Versioned V&V summary ↗Selected material, geometry, contact and execution records have distinct boundaries. Historical WP14 remains on hold and G03 remains failed.
Versioned evidence summary ↗32 of 46 historical combinations are qualified within scope; 14 remain experimental. Separate workflows do not alter these counts.
Versioned source registry ↗