[cmccomb]

Purpose

Run only the pinned Nonlinear_Hydro/ode45/Regular 0.000625 s source arm in a fresh R2025b MATLAB process. The earlier four-arm workflow suffered runner shutdown during this arm on three attempts; this checks whether memory accumulated from the three prior arms caused those interruptions.

Controls

• Pin the same WEC-Sim core and Applications revisions as Screen ellipsoid ode45 cases against refined full-duration source #655.
• Copy the published Regular application and change only simu.dt from 0.05 to 0.000625 s.
• Assert that simu.dt, simu.dtOut, simu.nonlinearDt, simu.cicDt, and Simulink MaxStep all equal 0.000625 s for the complete 150 s simulation.
• Upload the raw output history/settings plus generated BEM HDF5 and application STL for independent comparison against the previously measured first three source steps.

This diagnostic makes no status, badge, or published-input gate change. If it succeeds, the same isolated run can be applied to RegularCIC and incorporated into the source convergence screen. If it hits the same runner shutdown, the separately measured #659 output-grid A/B supplies evidence for testing a 0.005 s output clock while keeping the physical clocks at 0.000625 s.

Measured runner outcome

At exact head c8a4e3f41ad5a30bd77545944caf9817f4efe129, run 40105605489 began the all-clock 0.000625 s Regular simulation at 15:12:14 UTC on 2026-10-24. At 15:16:55 UTC the hosted runner reported a shutdown signal and cleaned up orphan processes. The run produced zero artifacts and never reached a MATLAB output or scientific gate. This fresh-process interruption shows that accumulation from the three earlier arms is not sufficient to explain the prior shutdowns; it does not measure physical parity. The failed run is retained as a diagnostic, and the next output-clock trial is isolated in a separate PR.

[cmccomb]

The isolated all-clock 0.000625 s attempt lost its hosted runner before producing a physical artifact; it is inconclusive. Merged #655 records that limit and uses an explicitly derived mixed-output finest source. This diagnostic PR is superseded.