tboudreaux/MeanField
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Record gravity relift work before descent rejection

Browse Source
This commit is contained in:
Emily Boudreaux
2026-09-17 12:17:10 -04:00
parent b9e6794999
commit c82e68dd7e
3 changed files with 31 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
experiments/GRAVITY_DIRECTION_DIAGNOSTICS_2026-09-17.md Normal file
View File

@@ -0,0 +1,26 @@
# Gravity work accounting on direction rejection
The combined polynomial-enthalpy and reduced-gravity strategy checks descent
twice: once when selecting the enthalpy repair, and again after lifting the
repaired retained direction into gravity coordinates. A successful gravity
relift performs real auxiliary work even when the second check rejects the
direction before line search.
Previously, the successful-relift diagnostics were copied only after the second
check. Its early rejection could therefore leave solve counts, iteration counts,
timings, and the last inner residual at their earlier linear-solve snapshot.
The snapshot now occurs immediately after successful direction capture and
before the final descent check. Linear-report coordinates retain their existing
meaning; this changes no tolerance, direction, state, or acceptance decision.
The immutable `b21fa40` rotating `positivity-reduced` artifact records 20 gravity
solves and a positive final merit slope. Those counters can omit a successful
last relift: they do not distinguish rejection during enthalpy repair from
rejection after relifting. Its `direction_repairs` count also includes a chosen
positive repair fraction whose candidate subsequently fails descent. Historical
artifacts must retain these qualifications rather than being retroactively
interpreted as complete work counters.
The regression is a repeat of that bounded combined rotating case, checking
unchanged accepted states and residual history while accounting for all
successful relift work. No extra numerical fixture or solver change is needed.

2
libmeanfield/interface/solver/newton_numerics.cppm
View File

@@ -72,6 +72,8 @@ export namespace mean_field::solver::nonlinear {
struct NumericalStrategyDiagnostics final {
int polynomialChecks{0};
// Nonzero repair proposals include directions later rejected by a
// descent check; this measures repair activity, not accepted steps.
int directionRepairs{0};
double lastRepairFraction{0.0};
double lastMeritDirectionalDerivative{0.0};

4
libmeanfield/interface/solver/stellar_newton_strategy.cppm
View File

@@ -440,6 +440,9 @@ namespace mean_field::solver::detail {
return false;
}
}
// A successful gravity relift still belongs in the work report if
// the following full-merit check rejects the repaired direction.
m_gravity.CopyDiagnostics(diagnostics);
if constexpr (altersDirection && std::same_as<Gravity, nonlinear::EliminatedGravity>) {
mfem::Vector action(state.Problem().EquationSize());
state.normalizedOperator->Mult(state.normalizedCorrection, action);
@@ -448,7 +451,6 @@ namespace mean_field::solver::detail {
if (!std::isfinite(diagnostics.lastMeritDirectionalDerivative) ||
diagnostics.lastMeritDirectionalDerivative >= 0.0) { return false; }
}
m_gravity.CopyDiagnostics(diagnostics);
return true;
}
auto PrepareTrial(State &state, const double alpha, nonlinear::NumericalStrategyDiagnostics &diagnostics) {