PSPASES Home
 Software
 Publications
 People
 Feedback
PSPASES (Parallel SPArse Symmetric dirEct Solver) is a high performance,
scalable, parallel, MPIbased library, intended for solving linear systems of
equations involving sparse symmetric positive definite matrices. The library
provides various
interfaces to solve the system using four phases of direct method of solution :
compute fillreducing ordering, perform symbolic
factorization, compute numerical factorization, and solve triangular systems of
equations. The library efficiently implements the scalable parallel
algorithms developed by the authors, to compute each of the phases
[GKK ,
JGKK ,
GGJKK ,
KK].
Features:
 High Performance Library. Solved a million equation system in 154 seconds on
Cray T3E with most computationally intensive phase clocking at 52 GFLOPS!
 Portable to most of today's parallel computers. Tested on IBM, Cray, and
SGI platforms.
 Entirely parallel and scalable code. Each of the four phases is
parallelized.
 Library functions can be called from both C and Fortran 90 codes, with
simple calling sequences.
 Memory requirements for the numerical factorization phase can be
preestimated.
 Concept of a communicator is used to enable solving multiple systems
with same sparsity structure, or same system for multiple sets of B.
 Command line interface provided, which supports four different matrix
formats, including the HarwellBoeing format.
PSPASES Software
Portability:
PSPASES can be used on any parallel computer or network of workstations
equipped with MPI and BLAS libraries, and Fortran 90 and C language compilers.
PSPASES functions can be called from both C and Fortran 90 programs.
It has been extensively tested on SGI Origin
2000, SGI PowerChallenge, Cray T3E,IBM SP, and network of IBM RS600
workstations.
Performance:
PSPASES could solve a system of 1 million equations with around 900 million
nonzeros in the filled matrix, in just 154 seconds on 256 processors of
Cray T3E. This time included times for all the phases of the solver. The
numerical factorization phase clocked a 52 GFLOPS performance, which in our
knowledge is the highest ever reported performance for a portable parallel
direct sparse solver. For more results depicting efficiency and scalability
of PSPASES, for various other matrices ranging
from few thousand to hundereds of thousands of equations, see this
Performance Table.
Obtaining the Software:
Get Latest Version

Changes from Previous Version

Get Older Version
A user's manual is supplied with the distribution. It explains
in detail, the formats of input and output parameters and the calling
sequences for various functions provided.
Here is a copy of the PSPASES user's manual
[PostScript 
PDF].
Read the
Install Instructions
first after downloading and uncompressing
the library. It has instructions for building and testing the library.
Read
Usage Notes
which may help you to get most out of PSPASES. It
answers some FAQs and explains various matrix
formats accepted by the command line interfaces provided. The
format specifications for two of the supported formats can be found in these
files :
HarwellBoeing format and
RutherfordBoeing format .
Related Software:
PSPASES uses ParMETIS and METIS as its default ordering libraries.
The current release of PSPASES supplies ParMETIS and METIS libraries
with the distribution. But, the versions supplied with PSPASES may not
always be the latest. For their latest versions and to find more about their
vast graphpartitioning related
functionality, refer to the METIS
site .
PSPASES uses the standard BLAS, LAPACK, and MPI library calls for its
functionality. Tuned versions of these libraries are recommended to get
good performance out of PSPASES.
Public domain
vanilla implementations of BLAS
and LAPACK routines are
available on the web.
A public domain implementation of MPI standard is available
at MPICH site .
Also, a faster version of PSPASES, with enhanced functionality, is available
for the IBM SP and RS6000 systems, as WSMP. It can solve symmetric positive
definite as well as indefinite systems. For more information, please visit
WSMP page
.
PSPASES Home
 Software
 Publications
 People
 Feedback
PSPASES related Publications
 [GKK]

Highly Scalable Parallel Algorithms for Sparse Matrix
Factorization (1995)
[PostScript]
Anshul Gupta, George Karypis, and Vipin Kumar.
 [JGKK]

A High Performance Two Dimensional Scalable Parallel Algorithm
for Solving Sparse Triangular Systems (1997)
[PostScript]
[PDF]
Mahesh Joshi, Anshul Gupta, George Karypis, and Vipin Kumar.
 [GGJKK]

PSPASES: An Efficient and Scalable Parallel Sparse Direct
Solver (1999)
[PostScript]
[PDF]
Anshul Gupta, Fred Gustavson, Mahesh Joshi, George Karypis,
and Vipin Kumar.
 [ParMETIS]

ParMETIS related Publications.
George Karypis and Vipin Kumar.
For more publications related to parallel sparse linear systems solutions
by Prof. Vipin Kumar's group, follow
this link
.
PSPASES Home
 Software
 Publications
 People
 Feedback
People associated with PSPASES

Anshul Gupta .
IBM Thomas J. Watson Research Center, Yorktown Heights, NY
10598.
Designed and developed the algorithm for parallel numerical factorization phase.
Helped in designing the triangular solution and symbolic factorization phases.

Mahesh Joshi .
Department of Computer Science, University of Minnesota,
Minneapolis, MN 55455.
Designed and developed algorithms for parallel symbolic factorization and
parallel triangular solution phases, and integrated all the four phases
of the solver to make it portable and useable in its current format.

George Karypis .
Department of Computer Science, University of Minnesota,
Minneapolis, MN 55455.
Designed and developed parallel algorithm for computing
fillreducing ordering.

Vipin Kumar .
Department of Computer Science, University of Minnesota,
Minneapolis, MN 55455.
Helped in designing algorithms for all the four phases of the solver.

Fred Gustavson.
IBM Thomas J. Watson Research Center, Yorktown Heights, NY
10598.
Supported the development of solver.
PSPASES Home
 Software
 Publications
 People
 Feedback
Feedback, Comments, Bug Report
We are eager to get positive as well as negative feedback from you,
about useability and quality of PSPASES.
Please contact Mahesh Joshi
(mjoshi@cs.umn.edu)
for any usage related questions or comments as well as to report
any abnormal behavior which could be possibly related to
bugs that are not discovered so far.
Back upto PSPASES Home
Get PSPASES 1.0.3
This Page Visited times.
Page Creator: Mahesh Joshi
Last modified:
Sun May 9 15:21:25 CDT 1999