The ARMAKES subdirectory contains sample files with machine specific information needed during the building of the ARPACK library. The BLAS and LAPACK subdirectories contain the necessary codes from the respective software libraries. The DOCUMENTS subdirectory contains files that have example templates of how to invoke the different computational modes offered by ARPACK. Example driver programs that illustrate all of the computational modes, data types and precisions may be found in the EXAMPLES directory. Programs for banded, complex, non-symmetric, symmetric eigenvalue problems and singular value decomposition may be found in the directories BAND, COMPLEX, NONSYM, SYM, SVD respectively. Look at the README files in each subdirectory for further information. The SRC subdirectory contains all the ARPACK source codes. The UTIL subdirectory contains the various utility routines needed for printing results and timing the execution of the ARPACK subroutines.
The archived library libarpack_<PLATFORM>.a is created upon completion of the installation instructions. Here, _<PLATFORM> denotes the environment where the ARPACK library is built. All of the subroutines other than those in the EXAMPLES directory are compiled and archived into libarpack_<PLATFORM>.a. The installer should be aware that the BLAS and LAPACK directories contain a subset of routines from these packages that will require an additional megabyte of memory once they are compiled and archived. If these packages are already available on your system, you may delete the BLAS and LAPACK directories provided here and point to the ones that are already installed on your system. This is easily done by modifying the file ARmake.inc as described in the README file in the top level directory ARPACK.
To get started, we recommend that the user enter the SIMPLE subdirectory and issue the commands
make dssimp ; dssimp > output
This will compile, link, and execute the dssimp program. dssimp is a sample driver for the reverse communication interface to the ARPACK routine dsaupd which finds a few eigenvalues and eigenvectors of a symmetric matrix.
This chapter discusses the use of dssimp for computing eigenvalues and eigenvectors of a symmetric matrix using the simplest computational mode. There are additional drivers available for all of the computational modes, data types, and precisions. These additional driver programs are in the EXAMPLES subdirectories. Each of them is self-contained and may be compiled and executed in a similar manner as described in the README files.
This dssimp driver should serve as template to enable a user to create a program to use dsaupd on a specific problem in the simplest computational mode. All of the driver programs in the various EXAMPLES subdirectories are intended to be used in this way. The simple programs have more extensive documentation to aid in the understanding and conversion but essentially the same principle and structure apply to all of the driver programs.
/ / / | \ \ \
ARMAKES BLAS DOCUMENTS | LAPACK SRC UTIL
/ / | \ \ \
SIMPLE BAND SVD SYM NONSYM COMPLEX