Crate ruvector_solver 

Iterative sparse linear solvers for the ruvector ecosystem.

This crate provides iterative methods for solving Ax = b where A is a sparse matrix stored in CSR format.

Available Solvers

Solver	Feature gate	Method
NeumannSolver	neumann	Neumann series x = sum (I-A)^k b

Example

use ruvector_solver::types::{ComputeBudget, CsrMatrix};
use ruvector_solver::neumann::NeumannSolver;
use ruvector_solver::traits::SolverEngine;

// Build a diagonally dominant 3x3 matrix (f32)
let matrix = CsrMatrix::<f32>::from_coo(3, 3, vec![
    (0, 0, 2.0_f32), (0, 1, -0.5_f32),
    (1, 0, -0.5_f32), (1, 1, 2.0_f32), (1, 2, -0.5_f32),
    (2, 1, -0.5_f32), (2, 2, 2.0_f32),
]);
let rhs = vec![1.0_f32, 0.0, 1.0];

let solver = NeumannSolver::new(1e-6, 500);
let result = solver.solve(&matrix, &rhs).unwrap();
assert!(result.residual_norm < 1e-4);

Modules

arena

Bump allocator for per-solve scratch space.

audit

Audit trail for solver invocations.

budget

Compute budget enforcement for solver operations.

cg

Conjugate Gradient solver for symmetric positive-definite systems.

error

Error types for the solver crate.

events

Event sourcing for solver operations.

forward_push

Forward Push solver for Personalized PageRank (Andersen-Chung-Lang).

neumann

Jacobi-preconditioned Neumann Series iterative solver.

router

Algorithm router and solver orchestrator.

simd

SIMD-accelerated sparse matrix-vector multiply.

traits

Solver trait hierarchy.

types

Core types for sparse linear solvers.

validation

Comprehensive input validation for solver operations.