Difficult problems and generalized problems will require one to use a shift-invert strategy that is based upon the use of a (sparse-direct) matrix factorization. These more sophisticated modes of operation are described in the next chapter. Example driver routines have been constructed for each problem type, computational mode, data type and precision. These drivers may be used as templates to construct a code for a specific application by substituting the appropriate data structures, matrix factorizations, solvers and matrix-vector products. An explanation of how to use these drivers as templates and how to modify them for your own use is given in Appendix A (Templates and Driver Routines). Each of the various drivers has been provided to address a typical situation arising from significant applications of eigenvalue calculations.

We begin with a description of the ARPACK directory structure. We then use one of the sample drivers to illustrate the use of ARPACK in the simplest mode of operation.