Struct M16TcDiff 

pub struct M16TcDiff {
    pub max_ulp: i32,
    pub over_budget: Vec<M16TcOutlier>,
    pub over_ulp_only: usize,
    pub sign_flips: usize,
    pub max_abs: f64,
    pub rel_rms: f64,
    pub rms: f64,
    pub row_rms: Vec<f64>,
}

2026-09-25: The result of comparing an m x n BF16 block against the scalar reference.

Fields

max_ulp: i32

2026-09-25: Largest ordinal BF16 ULP distance; sign flips are counted in sign_flips and excluded here.

over_budget: Vec<M16TcOutlier>

2026-09-25: Elements the full criterion rejected (ordinal budget and accumulation floor both exceeded).

over_ulp_only: usize

2026-09-25: Elements over the ordinal budget alone, whether or not the floor admitted them.

sign_flips: usize

max_abs: f64

rel_rms: f64

rms: f64

2026-09-25: RMS of the whole reference block. The floor uses the per-row values in row_rms, and this one only for a row that has none.

row_rms: Vec<f64>

2026-09-25: Per-row reference RMS, in row order: the scales the floor was evaluated at.

Trait Implementations

impl Clone for M16TcDiff

fn clone(&self) -> M16TcDiff

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for M16TcDiff

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

Formats the value using the given formatter.

impl Default for M16TcDiff

fn default() -> M16TcDiff

Returns the “default value” for a type.