Merged tests and examples

Merge pull request #7 from jaurentz/development

Development

Blocked matrix-vector multiplication

This release allows for blocked matrix-vector multiplication for added performance when using block Lanczos.

Introductory release

This release contains the basic framework for computing eigenvalues of large sparse symmetric matrices using Nvidia brand GPUs.