Linear Algebra for Structured Sparse Matrices

This is a Python library for representing structured, sparse matrices. The representation is compact and can be used in several ways:

• at run-time, to ensure the most efficient matrix-vector multiplication is carried out

• as a method to distributed sparse matrices across clusters when used in conjunction with PyTables

• as an intermediate representation for sparse linear algebra

It represents matrix structure constructively; it is not able to infer matrix structure given a sparse matrix. Instead, it requires that the user construct a sparse matrix using a set of primitive matrix structures.