sprs, a sparse matrix library written in Rust

sprs implements some sparse matrix data structures and linear algebra algorithms.

WARNING: experimental lib, neither API nor functionality stable

License

BSD license. See LICENSE.txt

Features

Structures ..........

• CSR/CSC matrix, able to operate on borrowed or owned data

Operations ..........

• sparse matrix vector product
• sparse matrix matrix product
• sparse matrix matrix addition

Algorithms ..........

• Outer iterator on compressed sparse matrices
• CSC/dense vector product
• CSR/dense vector product
• sparse Cholesky decomposition

TODO

Structures ..........

• CSC/CSR tests with more trickier shapes/data
• lower/upper triangular CSC/CSR matrices
• block-sparse matrices

Operations ..........

• rebind operations with the corresponding traits

Algorithms ..........

• tests on the outer iterator
• tests on the CSC/vector product
• sparse triangular solve
• sparse LU decomposition

Misc ....

• Python bindings

API guidelines

Each exposed functionality should, if deemed necessary for performance reasons, be exposed as a low-level C-style function (similar to BLAS API), and also exposed as a high level API with good defaults for the performance related arguments.