Struct LinearSolverSummary 

pub struct LinearSolverSummary {
    pub solver_name: String,
    pub n_factors: u64,
    pub n_pattern_reuse: u64,
    pub n_pattern_changes: u64,
    pub max_fill_ratio: Option<f64>,
    pub min_abs_pivot: Option<f64>,
    pub max_abs_pivot: Option<f64>,
    pub last_inertia: Option<(usize, usize, usize)>,
    pub last_nnz_a: Option<usize>,
    pub last_nnz_l: Option<usize>,
}

Aggregate stats accumulated over the lifetime of one linear-solver instance. All fields default to zero / None so a backend that declines to populate them still produces a valid summary.

Fields

solver_name: String

Short identifier of the backend that produced this summary: "feral", "ma57", etc. Empty for the Default value.

n_factors: u64

Number of factor() calls completed (including those that reused the cached symbolic factorisation).

n_pattern_reuse: u64

Of n_factors, how many reused the previous symbolic factorisation (sparsity pattern unchanged). Healthy IPM workloads expect this to dominate after the first iter.

n_pattern_changes: u64

Of n_factors, how many required a fresh symbolic factorisation (sparsity pattern changed). Inverse of n_pattern_reuse modulo the very first factor.

max_fill_ratio: Option<f64>

Maximum nnz(L) / nnz(A) observed across factors. Values much greater than ~10 on KKT-style systems indicate ordering trouble.

min_abs_pivot: Option<f64>

Minimum |pivot| observed across factors. Approaches the working-precision floor when the matrix is near-singular.

max_abs_pivot: Option<f64>

Maximum |pivot| observed across factors.

last_inertia: Option<(usize, usize, usize)>

Inertia of the final factorisation as (positive, negative, zero).

last_nnz_a: Option<usize>

nnz(A) of the final factorisation’s matrix.

last_nnz_l: Option<usize>

nnz(L) of the final factorisation.

Implementations

impl LinearSolverSummary

pub fn is_empty(&self) -> bool

Returns true if the summary carries no signal beyond the solver name — useful for “did we collect anything?” checks.

Trait Implementations

impl Clone for LinearSolverSummary

fn clone(&self) -> LinearSolverSummary

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LinearSolverSummary

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

Formats the value using the given formatter.

impl Default for LinearSolverSummary

fn default() -> LinearSolverSummary

Returns the “default value” for a type.