This is the release version as of 19-Sep-95
------------------------------------------
Initialized INFO to 0 in the function PDDGERES. Credits to 
Shennon Shen <xshen@TC.Cornell.EDU> for pointing this.

Purushotham Bangalore (puri@erc.msstate.edu; 601-325-4587}

-------------------------------------------------------------------------------

This is the release version as of 9-Dec-94
------------------------------------------
This directory contains files to run a double precision parallel dense
linear solve software. The original software available in the public
doamin is modified to make use of MPI message-passing primitives
instead of the BLACS calls.  At certain places, the BLACS interface is
retained with slight modification, and MPI primitives are used inside
the calls, whereas at certain places BLACS are totally replaced with
MPI calls. In order to help one understand the changes made, the
original BLACS calls are retained within comments. These programs have
been tested on network of workstations, Intel iPSC/860, and IBM SP1
machines. This is the first attempt to make use of MPI instead of
BLACS, hence there is still scope for further enhancements. Also the
original code was written for complex numbers, where as this code usese
double precision numbers. The estimation of one-norm condition number
is not included in this modified version of the code.

This is software provided for use with the book "Using MPI:  Portable
Parallel Programming with the Message-Passing Interface," by Gropp,
Lusk, Skjellum.  See page 149-152 (Section 6.4, entitled The LINPACK
Benchmark in MPI.)

Purushotham Bangalore (puri@erc.msstate.edu; 601-325-4587}

BRIEF INSTRUCTIONS

We include
   configure
   Makefile.in
which is a prototype makefile.  This is turned into a real makefile
by the program configure.

to allow configuration of this software on a wide variety of platforms.

		                         Makefile.in
		 configure ------------------>|
					      |
					  Makefile
					      |
		 make ----------------------->|
					      |
					   programs

Run
---
./configure (with OPTIONS) --- read the file configure (a Bourne shell script)
make

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
The following copyright notice that appears along with the original 
code is as follows: 

*
*     VERSION 2.0 DATED Aug. 1992
*     R. VAN DE GEIJN
*
*        This code resulted from a joint collaboration between  
*              Robert van de Geijn 
*     and the group for 'Research in Parallel Computational
*     Electromagnetics' at the Jet Propulsion Laboratory.  Futher 
*     information can be found in the paper,
*
*         T. Cwik, R. van de Geijn, and J. Patterson, 
*         "Application of Massively Parallel Computation to 
*          Integral Equation Models of Electromagnetic Scattering 
*          (Invited Paper)," J. Opt. Soc. Am. A, Vol. 11, No. 4, 
*          April 1994.
*
*     Please email your comments to the authors at
*
*        rvdg@cs.utexas.edu (Robert van de Geijn)
*        cwik@yosemite.jpl.nasa.gov (Tom Cwik)
*        jep@yosemite.jpl.nasa.gov  (Jean Patterson)
*
*     Use of this code must properly reference that paper.
*
*     The code is free, and comes with no implicit or explicit 
*     guarantees.  
*
*     All rights reserved
*     
