Module linear_algebra 

Linear Algebra Library - High-Performance Mathematical Computing

This module provides high-performance linear algebra operations leveraging Phase 2 enhancements:

• Hardware-Sympathetic Storage (ZNS) for zero-copy matrix operations
• NVMe Computational Storage (CSD) for hardware-accelerated computations
• Zero-Knowledge Semantic Proofs for privacy-preserving linear algebra
• Ambient Sub-Threshold Orchestration for mobile optimization

Re-exports

pub use crate::solvers::polynomial::Complex;

pub use crate::solvers::polynomial::QuadraticRoots;

pub use crate::solvers::linear_algebra::lu::Lu;

pub use crate::solvers::linear_algebra::svd::Svd;

pub use computation::*;

pub use core_types::*;

pub use optimization::*;

pub use performance::*;

pub use privacy::*;

pub use storage::*;

Modules

computation

core_types

optimization

performance

privacy

Privacy-preserving linear algebra.

storage

Structs

LinearAlgebraLibrary

Linear Algebra Library Manager

Functions

characteristic_polynomial

Characteristic polynomial — thin facade over the engine solvers::linear_algebra::spectral::characteristic_polynomial (Faddeev–LeVerrier).

determinant

Determinant of a row-major n×n matrix via LU decomposition — thin facade over the engine determinant. O(n³), numerically robust; returns 0.0 for a singular matrix.

eigen_symmetric

Eigen-decomposition of a SYMMETRIC row-major n×n matrix via cyclic Jacobi rotations. Returns (eigenvalues, eigenvectors) where eigenvectors is a row-major n×n matrix whose COLUMN j is the unit eigenvector for eigenvalues[j]. Errors if the input is not (within tolerance) symmetric.

eigenvalues_general

General (non-symmetric) eigenvalues — thin facade over the engine solvers::linear_algebra::spectral::eigenvalues_general.

lu_decompose

LU-decompose a row-major n×n matrix with partial pivoting — thin facade over the engine lu_decompose (maps the engine error to this lib’s error type).

polynomial_roots

Find all complex roots of a real polynomial — thin facade over the engine solvers::polynomial::polynomial_roots (Durand–Kerner).

solve_quadratic

Solve a·x² + b·x + c = 0 over the reals — thin facade over the engine solvers::polynomial::solve_quadratic (maps the engine error to this lib’s type).

svd

Singular value decomposition of a row-major m×n matrix — thin facade over the engine svd (maps the engine error to this lib’s error type). Singular values are returned in descending order.