Struct StepRankDiagnostics 

pub struct StepRankDiagnostics {
    pub pre_drift_ranks: Option<Vec<usize>>,
    pub post_drift_ranks: Option<Vec<usize>>,
    pub post_kick_ranks: Option<Vec<usize>>,
    pub post_final_ranks: Option<Vec<usize>>,
    pub poisson_rank_amplification: Option<f64>,
    pub advection_rank_amplification: Option<f64>,
    pub rank_growth_rate: Option<f64>,
    pub singular_value_spectrum: Option<Vec<Vec<f64>>>,
    pub rank_budget_fraction: f64,
}

Per-step rank diagnostics, populated by InstrumentedStrangSplitting. Fields are None when the representation is not HtTensor.

Fields

pre_drift_ranks: Option<Vec<usize>>

Ranks of each HT node before the first drift half-step.

post_drift_ranks: Option<Vec<usize>>

Ranks after drift(Δt/2).

post_kick_ranks: Option<Vec<usize>>

Ranks after the velocity kick (Poisson + advect_v).

post_final_ranks: Option<Vec<usize>>

Ranks after the second drift half-step (final state).

poisson_rank_amplification: Option<f64>

Poisson rank amplification: max(post_kick_rank) / max(post_drift_rank).

advection_rank_amplification: Option<f64>

Advection rank amplification: max(post_drift_rank) / max(pre_drift_rank).

rank_growth_rate: Option<f64>

Exponential fit to recent rank history (last 10 steps). Positive = rank growing, > 0.5 = doubling every 2 steps.

singular_value_spectrum: Option<Vec<Vec<f64>>>

Per-node singular values from the most recent HSVD truncation. Outer Vec = nodes (11 for 6D HT), inner Vec = singular values.

rank_budget_fraction: f64

Current max rank / max_rank budget (0.0 to 1.0).

Trait Implementations

impl Clone for StepRankDiagnostics

fn clone(&self) -> StepRankDiagnostics

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for StepRankDiagnostics

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for StepRankDiagnostics

fn default() -> StepRankDiagnostics

Returns the “default value” for a type.