KrylovMethods.jl

Released under the MIT License.

Simple and fast Julia implementation of Krylov subspace methods for linear systems.

Goals and Guidelines

The main goal of this package is to derive simple and fast implementation of the most useful Krylov subspace methods.

For the sake of simplicity aim to

• avoid declaration of our own types,
• avoid using greek letters in source codes,
• provide tested, readable and well-documented code.

Further goals are

• speed: where try to minimize allocation costs and maximize the use of BLAS routines,
• memory efficiency: storing temporary variables and re-allocations are avoided,
• generality: where possible complex systems are supported,
• reliability: unit tests are provided.