Struct ParallelEigenSolver 

pub struct ParallelEigenSolver { /* private fields */ }

Parallel Eigenvalue Decomposition

Implements parallel algorithms for eigenvalue decomposition of symmetric matrices commonly used in cross-decomposition methods. Uses divide-and-conquer approach for large matrices and multi-threading for improved performance.

Mathematical Background

For a symmetric matrix A, finds eigenvalues λ and eigenvectors v such that: A * v = λ * v

Uses parallel divide-and-conquer algorithm:

1. Decompose matrix into smaller blocks
2. Compute eigenvalues for each block in parallel
3. Merge results using parallel reduction

Examples

use sklears_cross_decomposition::ParallelEigenSolver;
use scirs2_core::ndarray::Array2;

let matrix = Array2::eye(100); // 100x100 identity matrix
let mut solver = ParallelEigenSolver::new()
    .n_threads(4)
    .block_size(25);

let (eigenvalues, eigenvectors) = solver.solve(&matrix).unwrap();

Implementations

impl ParallelEigenSolver

pub fn new() -> Self

Create a new parallel eigenvalue solver

pub fn with_thread_pool(self, pool: Arc<WorkStealingThreadPool>) -> Self

Use work-stealing thread pool for improved parallelism

pub fn n_threads(self, n_threads: usize) -> Self

Set number of threads

pub fn block_size(self, block_size: usize) -> Self

Set block size for divide-and-conquer

pub fn tolerance(self, tol: f64) -> Self

Set convergence tolerance

pub fn max_iterations(self, max_iter: usize) -> Self

Set maximum iterations

pub fn method(self, method: EigenMethod) -> Self

Set eigenvalue method

pub fn solve( &self, matrix: &Array2<f64>, ) -> Result<(Array1<f64>, Array2<f64>), SklearsError>

Solve eigenvalue problem

Trait Implementations

impl Clone for ParallelEigenSolver

fn clone(&self) -> ParallelEigenSolver

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ParallelEigenSolver

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

Formats the value using the given formatter.

impl Default for ParallelEigenSolver

fn default() -> Self

Returns the “default value” for a type.