klu

KLU is a software library for solving sparse linear systems using LU factorization. It is designed to handle large, sparse matrices efficiently, with particular emphasis on the kinds of systems that arise in circuit simulation. KLU provides both the factorization of a sparse matrix and solutions for multiple right-hand sides, aiming to minimize memory usage and computation time.

The library is implemented in C and offers a C API, with optional interfaces for other languages. It supports real-valued matrices and is built to work with sparse matrix formats common in numerical linear algebra, such as compressed sparse column storage. KLU emphasizes fill reduction and numerical robustness through its pivoting strategy, while keeping the implementation lightweight and portable across platforms. It also supports configurable index sizes (e.g., 32- or 64-bit indices) to accommodate large problems.

KLU is part of the SuiteSparse collection of sparse matrix algorithms and tools. It was developed by Tim Davis and collaborators to provide a fast, reliable solver tailored to circuit-oriented problems, but it can be used for a wide range of sparse LU factorization tasks. The library is released under an open-source license, allowing use, modification, and redistribution in both academic and commercial settings.

Availability and usage typically involve incorporating KLU into a software workflow that supplies a sparse matrix A and right-hand side vectors b. KLU computes an LU factorization of A and then solves Ax = b for one or more right-hand sides, potentially reusing the factorization for multiple solves to improve efficiency. Documentation and source code are distributed with the SuiteSparse project.