GUIDING QUESTION:
How can I exploit matrix structure in solving linear systems?
In many cases we can exploit structure in the coefficent matrix to develop faster and more stable numerical methods, and to guarantee the convergence of our methods.