sparsekit

sparsekit, a Fortran90 code which carries out a number of operations on sparse matrices, particularly conversion between various sparse formats.

sparsekit() can manipulate sparse matrices in a variety of formats, and can convert from one to another. For example, a matrix can be converted from the generalized diagonal format used by ELLPACK() and ITPACK() to the format used by the Harwell-Boeing Sparse Matrix Collection or into LINPACK() banded format.

Utilities available include converting data structures, printing simple statistics on a matrix, plotting a matrix profile, performing basic linear algebra operations (similar to the BLAS for dense matrix), and so on.

Matrix formats that are recognized include:

• BND, the LINPACK() format for general banded matrices.
• BSR, block row sparse format.
• COO, coordinate format.
• CSC, compressed sparse column format.
• CSR, compressed sparse row format.
• DIA, the diagonal sparse matrix format (NOT a diagonal matrix!).
• DIAG, a diagonal matrix, stored as a vector.
• DNS, dense storage, also called full storage.
• ELL, the format used by ELLPACK() and ITPACK().
• HB, Harwell-Boeing format. (Actually, CSR format plus auxilliary data)
• JAD, the jagged diagonal format.
• LNK, linked storage format.
• MSC, modified sparse column format.
• MSR, modified sparse row format.
• SSK, Symmetric skyline format.
• SSR, Symmetric sparse row format.

Licensing:

The information on this web page is distributed under the MIT license.

Languages:

sparsekit is available in a Fortran90 version.

Related Data and Programs:

sparsekit_test

dlap, a Fortran90 code which solves sparse linear systems.

hb_io, a Fortran90 code which reads and writes sparse linear systems stored in the Harwell-Boeing Sparse Matrix format.

hb_to_st, a Fortran77 code which converts the sparse matrix information stored in a Harwell-Boeing file into a sparse triplet file.

MGMRES, a Fortran90 code which applies the restarted GMRES algorithm to solve a sparse linear system.

mm_io, a Fortran90 code which reads and writes sparse linear systems stored in the Matrix Market format.

sparse_ccs, a data directory which contains a description and examples of the CCs format, ("compressed column sparse") for storing a sparse matrix, including a way to write the matrix as a set of three files.

sparse_crs, a data directory which contains a description and examples of the CRS format, ("compressed row sparse") for storing a sparse matrix, including a way to write the matrix as a set of three files.

sparsepak, a Fortran90 code which reorders and solves sparse linear systems.

Reference:

1. Efstratios Gallopoulos, Youcef Saad,
   Efficient solution of parabolic equations by Krylov approximation methods,
   RIACS Technical Report, 90-14.
2. Noborou Kikuchi,
   Finite element methods in mechanics,
   Cambridge University Press, 1986.
3. David Kincaid, Thomas Oppe, John Respess, David Young,
   ITPACKV 2C User's Guide,
   Technical Report CNA-191.
   Center for Numerical Analysis,
   University of Texas at Austin, 1984.
4. Donald Knuth,
   The Art of Computer Programming,
   Volume 3: Sorting and Searching,
   Addison-Wesley, 1973.
5. Ole Osterby, Zahari Zlatev,
   Direct Methods for Sparse Matrices,
   Springer-Verlag 1983.
6. Youcef Saad,
   Sparsekit: a basic tool kit for sparse matrix computations,
   Technical Report, Computer Science Department,
   University of Minnesota, June 1994.
7. Youcef Saad,
   Analysis of some Krylov subspace approximations to the matrix exponential operator,
   RIACS Technical Report, 90-14.
8. Yousef Saad,
   Iterative Methods for Sparse Linear Systems,
   Second Edition,
   SIAM, 2003,
   ISBN: 0898715342.
9. Zahari Zlatev, Kjeld Schaumburg, Jerzy Wasniewski,
   A testing Scheme for Subroutines Solving Large Linear Problems,
   Computers and Chemistry,
   Volume 5, Number 2-3, pages 91-100, 1981.

Source Code:

• sparsekit.f90, the source code.
• sparsekit.sh, compiles the source code.