Struct OptimizationStats 

pub struct OptimizationStats {
    pub ops_before: usize,
    pub ops_after: usize,
    pub bindings_before: usize,
    pub bindings_after: usize,
    pub literals_before: usize,
    pub literals_after: usize,
    pub iterations: usize,
    pub converged: bool,
}

Per-pipeline-run statistics. Useful for benchmarks, regression diagnostics, and --verbose emit modes.

Fields

ops_before: usize

Top-level body op count before optimization.

ops_after: usize

Top-level body op count after optimization.

bindings_before: usize

Number of bindings before. After equals bindings_before minus what drop_unused_bindings stripped.

bindings_after: usize

literals_before: usize

Number of literals in the top-level body’s pool, before/after.

literals_after: usize

iterations: usize

How many run_all_once iterations actually fired (1 to RUN_ALL_MAX_ITERS). 1 means the pipeline converged on the first pass; higher means later-pass changes exposed work for earlier passes and the fixed-point loop kicked in.

converged: bool

True iff the fixed-point converged within the cap. False means the pipeline stopped mid-way; output is still valid (every individual pass is total) but may have residual optimization opportunities.

Implementations

impl OptimizationStats

pub fn ops_eliminated(&self) -> usize

Total ops eliminated (top-level body only). Saturating, so the rare case where output exceeds input (e.g. loop_unroll inlining) returns 0 rather than wrapping.

pub fn bindings_dropped(&self) -> usize

pub fn is_no_op(&self) -> bool

True iff the pipeline made no change at all - no ops eliminated, no bindings dropped, no literals dropped, AND op count is stable. The kernel was either already optimal or out of the pipeline’s reach. Useful for tooling that wants to skip emit re-runs when nothing changed.

pub fn off_graph_dropped(&self) -> usize

Total off-graph data dropped (bindings + literals). Useful as a single-number “how much cleanup did the tail of the pipeline do?” signal.

pub fn merge(&mut self, other: OptimizationStats)

Merge another OptimizationStats into a running aggregate. Adds counts and ORs the converged flag (any non-converged run → aggregate is non-converged). iterations accumulates. Useful for tooling that runs run_all_with_stats over a corpus of N kernels and wants a single rolled-up summary.

pub fn zero() -> Self

Identity element for Self::merge. Useful as the seed of a fold over a corpus.

pub fn format_short(&self) -> String

One-line human-readable summary suitable for log lines. Format: "ops X→Y (-N), bindings A→B (-M), iters K (converged|stopped)". Mirrors the format_short pattern on the audit reports.

Trait Implementations

impl Clone for OptimizationStats

fn clone(&self) -> OptimizationStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for OptimizationStats

impl Debug for OptimizationStats

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for OptimizationStats

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for OptimizationStats

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

Formats the value using the given formatter.

impl Eq for OptimizationStats

impl PartialEq for OptimizationStats

fn eq(&self, other: &OptimizationStats) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for OptimizationStats

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for OptimizationStats