Struct ReedSolomonCode

pub struct ReedSolomonCode<F> { /* private fields */ }

Reed–Solomon codes over binary fields.

The Reed–Solomon code admits an efficient encoding algorithm over binary fields due to LCH14. The additive NTT encoding algorithm encodes messages interpreted as the coefficients of a polynomial in a non-standard, novel polynomial basis and the codewords are the polynomial evaluations over a linear subspace of the field. See the [binius-math] crate for more details.

Implementations

impl<F> ReedSolomonCode<F>

pub fn subspace(&self) -> &BinarySubspace<F>

impl<F> ReedSolomonCode<F>

pub fn log_inv_rate(&self) -> usize

impl<F: BinaryField> ReedSolomonCode<F>

pub fn new(log_dimension: usize, log_inv_rate: usize) -> Result<Self, Error>

pub fn with_ntt_subspace( ntt: &impl AdditiveNTT<Field = F>, log_dimension: usize, log_inv_rate: usize, ) -> Result<Self, Error>

pub fn with_domain_context_subspace( domain_context: &impl DomainContext<Field = F>, log_dimension: usize, log_inv_rate: usize, ) -> Result<Self, Error>

pub fn with_subspace( subspace: BinarySubspace<F>, log_dimension: usize, log_inv_rate: usize, ) -> Result<Self, Error>

pub const fn dim(&self) -> usize

The dimension.

pub const fn log_dim(&self) -> usize

pub const fn log_len(&self) -> usize

pub const fn len(&self) -> usize

The block length.

pub const fn inv_rate(&self) -> usize

The reciprocal of the rate, ie. self.len() / self.dim().

pub fn encode_batch_inplace<P: PackedField<Scalar = F>, NTT: AdditiveNTT<Field = F> + Sync>( &self, ntt: &NTT, code: &mut [P], log_batch_size: usize, ) -> Result<(), Error>

Encode a batch of interleaved messages in-place in a provided buffer.

The message symbols are interleaved in the buffer, which improves the cache-efficiency of the encoding procedure. The interleaved codeword is stored in the buffer when the method completes.

Throws

• If the code buffer does not have capacity for len() << log_batch_size field elements.

pub fn encode_batch<P: PackedField<Scalar = F>, NTT: AdditiveNTT<Field = F> + Sync>( &self, ntt: &NTT, data: &[P], output: &mut [MaybeUninit<P>], log_batch_size: usize, ) -> Result<(), Error>

Encode a batch of interleaved messages into a provided output buffer.

This function encodes multiple messages in parallel by copying the input data to all chunks of the output buffer and then applying the NTT transformation. The messages are interleaved in both the input and output buffers, which improves cache efficiency.

Preconditions

• The input data must contain exactly dim() << log_batch_size field elements.
• The output buffer must have capacity for len() << log_batch_size field elements.

Postconditions

• On success, all elements in the output buffer are initialized with the encoded codeword.

Throws

• Error::EncoderSubspaceMismatch if the NTT subspace doesn’t match the code’s subspace.
• Error::Math if the output buffer has incorrect dimensions.

pub fn encode_ext_batch_inplace<PE: PackedExtension<F>, NTT: AdditiveNTT<Field = F> + Sync>( &self, ntt: &NTT, code: &mut [PE], log_batch_size: usize, ) -> Result<(), Error>

Encode a batch of interleaved messages of extension field elements in-place in a provided buffer.

A linear code can be naturally extended to a code over extension fields by encoding each dimension of the extension as a vector-space separately.

Preconditions

• PE::Scalar::DEGREE must be a power of two.

Throws

• If the code buffer does not have capacity for len() << log_batch_size field elements.

pub fn encode_ext_batch<PE: PackedExtension<F>, NTT: AdditiveNTT<Field = F> + Sync>( &self, ntt: &NTT, data: &[PE], output: &mut [MaybeUninit<PE>], log_batch_size: usize, ) -> Result<(), Error>

Encode a batch of interleaved messages of extension field elements into a provided buffer.

A linear code can be naturally extended to a code over extension fields by encoding each dimension of the extension as a vector-space separately.

Preconditions

• PE::Scalar::DEGREE must be a power of two.
• The input data must contain exactly dim() << log_batch_size field elements.
• The output buffer must have capacity for len() << log_batch_size field elements.

Postconditions

• On success, all elements in the output buffer are initialized with the encoded codeword.

Throws

• Error::EncoderSubspaceMismatch if the NTT subspace doesn’t match the code’s subspace.
• Error::Math if the output buffer has incorrect dimensions.

Trait Implementations

impl<F: Clone> Clone for ReedSolomonCode<F>

fn clone(&self) -> ReedSolomonCode<F>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<F: Debug> Debug for ReedSolomonCode<F>

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

Formats the value using the given formatter.