Struct EffectEstimate 

pub struct EffectEstimate {
    pub shift_ns: f64,
    pub tail_ns: f64,
    pub credible_interval_ns: (f64, f64),
    pub pattern: EffectPattern,
    pub interpretation_caveat: Option<String>,
}

Estimated timing effect decomposed into shift and tail components.

The effect is decomposed using a 2-component linear model:

• Shift: Uniform timing difference across all quantiles (e.g., different code path)
• Tail: Upper quantiles shift more than lower (e.g., cache misses)

See spec Section 2.5 (Bayesian Inference).

Fields

shift_ns: f64

Uniform shift in nanoseconds.

Positive value means the sample class is slower (timing leak detected). Negative value means the sample class is faster (no leak, or unusual). This captures effects like branch timing where all operations take a fixed additional time.

tail_ns: f64

Tail effect in nanoseconds.

Positive value means the sample class has a heavier upper tail. This captures effects like cache misses that occur probabilistically.

credible_interval_ns: (f64, f64)

95% credible interval for the total effect magnitude in nanoseconds.

This is a Bayesian credible interval, not a frequentist confidence interval. There is a 95% posterior probability that the true effect lies within this range.

pattern: EffectPattern

Classification of the dominant effect pattern.

interpretation_caveat: Option<String>

When Some, the (μ, τ) decomposition may be unreliable.

This is set when model fit is poor (Q > q_thresh), indicating that the observed quantile pattern is not well-explained by the shift+tail basis. The shift_ns and tail_ns values should be interpreted with caution.

Implementations

impl EffectEstimate

pub fn total_effect_ns(&self) -> f64

Compute the total effect magnitude (L2 norm of shift and tail).

pub fn is_negligible(&self, threshold_ns: f64) -> bool

Check if the effect is negligible (both components near zero).

Trait Implementations

impl Clone for EffectEstimate

fn clone(&self) -> EffectEstimate

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for EffectEstimate

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

Formats the value using the given formatter.

impl Default for EffectEstimate

fn default() -> EffectEstimate

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for EffectEstimate

fn deserialize<__D>( __deserializer: __D, ) -> Result<EffectEstimate, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for EffectEstimate

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

where __S: Serializer,

Serialize this value into the given Serde serializer.