Struct MultilinearExtension

pub struct MultilinearExtension<P: PackedField, Data: Deref<Target = [P]> = Vec<P>> { /* private fields */ }

A multilinear polynomial represented by its evaluations over the boolean hypercube.

This polynomial can also be viewed as the multilinear extension of the slice of hypercube evaluations. The evaluation data may be either a borrowed or owned slice.

The packed field width must be a power of two.

Implementations

impl<P, Data> MultilinearExtension<P, Data>

where P: PackedField, Data: Deref<Target = [P]>,

pub fn specialize<PE>(self) -> MLEEmbeddingAdapter<P, PE, Data>

where PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>,

impl<'a, P, Data> MultilinearExtension<P, Data>

where P: PackedField, Data: Deref<Target = [P]> + Send + Sync + Debug + 'a,

pub fn specialize_arc_dyn<PE>( self, ) -> Arc<dyn MultilinearPoly<PE> + Send + Sync + 'a>

where PE: PackedField + RepackedExtension<P>, PE::Scalar: ExtensionField<P::Scalar>,

impl<P: PackedField> MultilinearExtension<P>

pub fn zeros(n_vars: usize) -> Result<Self, Error>

pub fn from_values(v: Vec<P>) -> Result<Self, Error>

impl<P: PackedField, Data: Deref<Target = [P]>> MultilinearExtension<P, Data>

pub fn from_values_generic(v: Data) -> Result<Self, Error>

pub fn new(n_vars: usize, v: Data) -> Result<Self, Error>

impl<U, F, Data> MultilinearExtension<PackedType<U, F>, PackingDeref<U, F, Data>>

where U: UnderlierType + PackScalar<F>, F: Field, Data: Deref<Target = [U]>,

pub fn from_underliers(v: Data) -> Result<Self, Error>

impl<'a, P: PackedField> MultilinearExtension<P, &'a [P]>

pub fn from_values_slice(v: &'a [P]) -> Result<Self, Error>

impl<P: PackedField, Data: Deref<Target = [P]>> MultilinearExtension<P, Data>

pub fn n_vars(&self) -> usize

pub fn size(&self) -> usize

pub fn evals(&self) -> &[P]

pub fn to_ref(&self) -> MultilinearExtension<P, &[P]>

pub fn packed_evaluate_on_hypercube(&self, index: usize) -> Result<P, Error>

Get the evaluations of the polynomial on a subcube of the hypercube of size equal to the packing width.

Arguments

• index - The index of the subcube

pub fn evaluate_on_hypercube(&self, index: usize) -> Result<P::Scalar, Error>

impl<P, Data> MultilinearExtension<P, Data>

where P: PackedField, Data: Deref<Target = [P]> + Send + Sync,

pub fn evaluate<'a, FE, PE>( &self, query: impl Into<MultilinearQueryRef<'a, PE>>, ) -> Result<FE, Error>

where FE: ExtensionField<P::Scalar>, PE: PackedField<Scalar = FE>,

pub fn evaluate_partial_high<'a, PE>( &self, query: impl Into<MultilinearQueryRef<'a, PE>>, ) -> Result<MultilinearExtension<PE>, Error>

where PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>,

Partially evaluate the polynomial with assignment to the high-indexed variables.

The polynomial is multilinear with $\mu$ variables, $p(X_0, …, X_{\mu - 1}$. Given a query vector of length $k$ representing $(z_{\mu - k + 1}, …, z_{\mu - 1})$, this returns the multilinear polynomial with $\mu - k$ variables, $p(X_0, …, X_{\mu - k}, z_{\mu - k + 1}, …, z_{\mu - 1})$.

REQUIRES: the size of the resulting polynomial must have a length which is a multiple of PE::WIDTH, i.e. 2^(\mu - k) \geq PE::WIDTH, since WIDTH is power of two

pub fn evaluate_partial_low<'a, PE>( &self, query: impl Into<MultilinearQueryRef<'a, PE>>, ) -> Result<MultilinearExtension<PE>, Error>

where PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>,

Partially evaluate the polynomial with assignment to the low-indexed variables.

The polynomial is multilinear with $\mu$ variables, $p(X_0, …, X_{\mu-1}$. Given a query vector of length $k$ representing $(z_0, …, z_{k-1})$, this returns the multilinear polynomial with $\mu - k$ variables, $p(z_0, …, z_{k-1}, X_k, …, X_{\mu - 1})$.

REQUIRES: the size of the resulting polynomial must have a length which is a multiple of P::WIDTH, i.e. 2^(\mu - k) \geq P::WIDTH, since WIDTH is power of two

pub fn evaluate_partial_low_into<PE>( &self, query: MultilinearQueryRef<'_, PE>, out: &mut [PE], ) -> Result<(), Error>

where PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>,

Partially evaluate the polynomial with assignment to the low-indexed variables.

The polynomial is multilinear with $\mu$ variables, $p(X_0, …, X_{\mu-1}$. Given a query vector of length $k$ representing $(z_0, …, z_{k-1})$, this returns the multilinear polynomial with $\mu - k$ variables, $p(z_0, …, z_{k-1}, X_k, …, X_{\mu - 1})$.

REQUIRES: the size of the resulting polynomial must have a length which is a multiple of P::WIDTH, i.e. 2^(\mu - k) \geq P::WIDTH, since WIDTH is power of two

impl<F: Field + AsSinglePacked, Data: Deref<Target = [F]>> MultilinearExtension<F, Data>

pub fn to_single_packed(self) -> MultilinearExtension<F::Packed>

Convert MultilinearExtension over a scalar to a MultilinearExtension over a packed field with single element.

Trait Implementations

impl<P, Data> AsRef<MultilinearExtension<P, Data>> for MLEDirectAdapter<P, Data>

where P: PackedField, Data: Deref<Target = [P]>,

fn as_ref(&self) -> &MultilinearExtension<P, Data>

Converts this type into a shared reference of the (usually inferred) input type.

impl<P, PE, Data> AsRef<MultilinearExtension<P, Data>> for MLEEmbeddingAdapter<P, PE, Data>

where P: PackedField, PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>, Data: Deref<Target = [P]>,

fn as_ref(&self) -> &MultilinearExtension<P, Data>

Converts this type into a shared reference of the (usually inferred) input type.

impl<P: Clone + PackedField, Data: Clone + Deref<Target = [P]>> Clone for MultilinearExtension<P, Data>

fn clone(&self) -> MultilinearExtension<P, Data>

Returns a copy of the value.

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

Performs copy-assignment from source.

impl<P: Debug + PackedField, Data: Debug + Deref<Target = [P]>> Debug for MultilinearExtension<P, Data>

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

Formats the value using the given formatter.

impl<P, Data> From<MultilinearExtension<P, Data>> for MLEDirectAdapter<P, Data>

where P: PackedField, Data: Deref<Target = [P]>,

fn from(inner: MultilinearExtension<P, Data>) -> Self

Converts to this type from the input type.

impl<P, PE, Data> From<MultilinearExtension<P, Data>> for MLEEmbeddingAdapter<P, PE, Data>

where P: PackedField, PE: PackedField, PE::Scalar: ExtensionField<P::Scalar>, Data: Deref<Target = [P]>,

fn from(inner: MultilinearExtension<P, Data>) -> Self

Converts to this type from the input type.

impl<P: PartialEq + PackedField, Data: PartialEq + Deref<Target = [P]>> PartialEq for MultilinearExtension<P, Data>

fn eq(&self, other: &MultilinearExtension<P, Data>) -> bool

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl<P: Eq + PackedField, Data: Eq + Deref<Target = [P]>> Eq for MultilinearExtension<P, Data>

impl<P: PackedField, Data: Deref<Target = [P]>> StructuralPartialEq for MultilinearExtension<P, Data>