Monday, 31 March 2014


This blog will contain status updates on my Google Summer of Code
(GSoC) project "Parallel Computation of Matrix Invariant Factors in

My plan is to implement an algorithm for determining the invariant
factors of a matrix by the Block-Wiedemann algorithm. Some key

  - All of this is being done in LinBox, so a lot of
    the scaffolding is already in place.

  - The input matrix A will be big, sparse and populated by elements
    drawn from a finite field.

  - The real focus of this project is not just to compute invariant
    factors, but to do so fast. To help that along I'm going to use
    shared memory parallelism via OpenMP. Making things work
    efficiently in parallel is almost always harder than you expect,
    so this will most likely be the most time-consuming part of the

More specifics can be found in my proposal here, and new posts will
appear here as the project progresses.

No comments:

Post a Comment