[us-commits] [ehb54/ultrascan3] f3e555: somo/grpy: name the worst observable, and stop ann...

[emre brookes]

  Branch: refs/heads/ehb54-issue-984
  Home:   https://github.com/ehb54/ultrascan3
  Commit: f3e555a32854d3ee39df4d5df0901640384580c2
      https://github.com/ehb54/ultrascan3/commit/f3e555a32854d3ee39df4d5df0901640384580c2
  Author: ehb54 <brookes at uthscsa.edu>
  Date:   2026-08-07 (Fri, 07 Aug 2026)

  Changed paths:
    M us_somo/develop/src/us_hydrodyn_grpy.cpp

  Log Message:
  -----------
  somo/grpy: name the worst observable, and stop announcing an inert cap

Both found by a real capped run (3GUT vdW, GRPY_SHELL_MAX_BEADS=2000).

The result line quoted err_max alone: "estimated error 3.85%". That is the
MAX over the requested observables, and intrinsic viscosity runs ~3.3x the
error of D_t at equal reduction, so on an unconverged run the single number
quoted is essentially always the viscosity's. Reading it as the accuracy of
the whole calculation makes D_t look several times worse than it is -- in that
run, 1416 of 11328 beads kept implies a D_t bar near 1.15%, not 3.85%.
ShellReport::worst was already computed for exactly this and had never been
shown; it is now named, with a following line pointing at the per-observable
estimates in the results file. Suppressed on an exact (unreduced) result,
where every bar is zero and "worst" means nothing.

The cap override also announced itself when shell reduction was OFF, where it
bounds nothing: the same run printed "bead cap OVERRIDDEN to 2000" and then
refused the model on memory, implying the 2000 had caused the refusal when it
was irrelevant. It now says so when inert, rather than going silent -- a stale
environment variable should still be visible.

Fixes ehb54/ultrascan-tickets#984

Co-Authored-By: Claude Opus 5 <noreply at anthropic.com>



To unsubscribe from these emails, change your notification settings at https://github.com/ehb54/ultrascan3/settings/notifications