Struct MatchingDecoder 

pub struct MatchingDecoder { /* private fields */ }

Exact minimum-weight perfect matching decoder compiled from a graphlike detector error model.

The decoding graph is the union-find decoder’s: detectors are vertices, a two-detector mechanism an internal edge and a one-detector mechanism a boundary edge, each weighted ln((1-p)/p) clamped at zero, with mechanisms sharing one detector set collapsed to the most probable. Weights are quantized to 2^-23 of the largest edge weight, and the matching is exact for the quantized weights: each shot’s correction has the minimum total weight of any edge set whose boundary is the shot’s defects, the open boundary absorbing any number of chain ends. Decoding uses no randomness, so equal inputs give equal outputs on any thread count. The algorithm is described in the decoding section of docs/architecture/qec-programs.md.

Implementations

impl MatchingDecoder

pub fn from_model(model: &DetectorErrorModel) -> Result<Self>

Compile a decoder from a graphlike detector error model.

Errors

As super::UnionFindDecoder::from_model: a mechanism flipping more than two detectors is rejected with a pointer to DetectorErrorModel::decompose_graphlike, and probabilities must lie in [0, 1).

pub fn num_detectors(&self) -> usize

pub fn num_observables(&self) -> usize

pub fn decode_packed(&self, detectors: &PackedShots) -> Result<PackedShots>

Decode packed detector samples into predicted observable flips.

Layouts and bit order as super::UnionFindDecoder::decode_packed.

Errors

The input measurement count must equal the model’s detector count, and a shot with an odd number of defects in a detector component without a boundary edge rejects the batch, naming the first such shot.

pub fn logical_error_rate( &self, detectors: &PackedShots, observables: &PackedShots, ) -> Result<f64>

Decode detectors and return the fraction of shots whose predicted flips differ from observables in any observable; 0.0 for no shots.

Errors

As Self::decode_packed, and observables must hold one bit per observable for the same shot count.

Trait Implementations

impl Clone for MatchingDecoder

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MatchingDecoder

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

Formats the value using the given formatter.