From num-alg-grp.co.uk!richard Fri Jun 12 13:34:41 WET 1992
From: Richard Brankin <richard@num-alg-grp.co.uk>
Message-Id: <18640.9206121234@nagds1.nag.co.uk>
Subject: netlib contribution
To: ehg@research.att.com, trh@ukc.ac.uk
Date: Fri, 12 Jun 92 13:34:41 WET DST
To: "Eric Grosse - netlib overlord Hopkins - netlib overlord"@uk.ac.uks
PP_warning: Parse error in original version of preceding line

I would like to contribute "rksuite" to netlib. Unforunately I can't send 
you send the tar file through our mailer due to it's size and the
sys admin chap isn't around. A compressed tar file is available via
anonymous ftp from "seas.smu.edu" in directory pub - rksuite.tar.Z.

So either 1. you could pick it up there
      or  2. I can send you rksuite in pieces
      or  3. you give me ftp-write access to an account and I'll 
             ftp it to you

The only thing I wish changed in the version of rksuite available at SMU
is the read.me file - revised version attached below my sig.

Here is some information for the netlib indices

lib     rksuite
for     solution of initial value problem in ordinary differential equations
#       A suite of codes for solving IVPs in ODEs. A choice of RK methods
#       is available. Includes an error assessment facility and a sophisticated
#       stiffness checker. Template programs and example results provided.
alg     Runge-Kutta
ref     RKSUITE, Softreport 92-S1, Dept of Math, SMU, Dallas, Texas
by      R.W. Brankin (NAG), I. Gladwell and L.F. Shampine (SMU)
lang    Fortran 77
prec    double

Please let me know  if there is any other information you require or any
modifications you would like me to make.

Regards, Richard.
-- 
\\______Richard Brankin______na.brankin@na-net.ornl.gov______//
//          NAG Ltd.            richard@nag.co.uk            \\

      **********************RKSUITE read.me file *************************
                   ****************************************


                  RKSUITE Release 1.0  November 1991

                             written by

          R.W. Brankin (*),  I. Gladwell(**),  and  L.F. Shampine (**)

 (*)  Numerical Algorithms Group Ltd.
      Wilkinson House
      Jordan Hill Road
      Oxford OX2 8DR
      U.K.
      email: richard@nag.co.uk
             na.brankin@na-net.ornl.gov
      International phone: + 44 865 511245
      International fax: + 44 865 310139

 (**) Department of Mathematics
      Southern Methodist University
      Dallas, Texas 75275
      U.S.A.
      email: h5nr1001@vm.cis.smu.edu
      U.S. phone: (214) 692-2542
      U.S. fax: (214) 692-4138


RKSUITE is a suite of codes based on Runge-Kutta methods for the numerical
solution of the initial value problem for a first order system of ordinary
differential equations.  It is the result of decades of research and
development of such methods and software by the authors.  It supersedes
some very widely used codes written by the authors and their coauthors,
namely, the RKF45 code that is available in several books and its descendant
DDERKF in the SLATEC library, and D02PAF and the associated codes in the NAG
Fortran library.

RKSUITE is being made available free of charge to the scientific community as
a public service.  It is expected that anyone making substantial use of the
software will acknowledge this use, and in particular, give a proper citation
in any publications resulting from this use.  A suitable reference is:

           R.W. Brankin, I. Gladwell, and L.F. Shampine, RKSUITE: a suite
           of Runge-Kutta codes for the initial value problem for ODEs,
           Softreport 92-S1, Department of Mathematics, Southern Methodist 
           University, Dallas, Texas, U.S.A, 1992.
         
The authors have tested the codes on a variety of problems, computers, and
compilers.  The codes are believed to perform correctly on problems for
which they were designed.  Of course, errors are possible in a software 
project of this size.  The authors assume no responsibility for the 
consequences of errors resulting from the use of this free software.  They
would greatly appreciate notification of unsatisfactory performance that 
might indicate the presence of a bug.  Constructive criticism would also be
much appreciated.

Future releases are planned that will add capabilities to the suite and 
correct any errors that might be discovered.  

                     ****************************************
         ************************************************************
                     ****************************************

On the distribution diskette are:

read.me            -  this file.

DOCS               -  a DIRECTORY containing two documentation files

    rksuite.doc    -  the basic documentation that explains how to use the 
                      suite of codes to solve initial value problems. 
                      
                      ============================================
                      ============================================
                      YOU SHOULD READ THE DOCUMENTATION CAREFULLY.
                      ============================================
                      ============================================

    details.doc    -  provides some details about how the codes are organized.
                      In particular, there is a list showing for each routine
                      which routines it calls and which call it directly. Also,
                      routines communicate with each other by means of labelled
                      COMMON blocks.  They are described and details of their 
                      usage by the routines are given.

CODES              -  a DIRECTORY containing one file of subprograms

    rksuite.for    -  source code for the FORTRAN subprograms that constitute
                      the suite.  Each user-callable subprogram has a prolog 
                      that provides a "Brief Reminder" of the meaning, type,
                      and size requirements of the arguments.                

TMPLS              -  a DIRECTORY that contains three templates illustrating
                      how to use the suite to solve typical problems.  Each
                      template comes in two variants.  An output file for
                      each template showing the results of the use of the
                      template in both its variants is provided.  Other
                      choices of methods are possible in some templates; 
                      output using them is included in the output file.
                      That is, there are nine files in total (six with 
                      extension .for and three with extension .out).

                      ===================================================
                      ===================================================
                      THE EASIEST WAY TO SOLVE A PROBLEM IS OFTEN TO EDIT 
                      ONE OF THE TEMPLATES PROVIDED IN THIS DIRECTORY.
                      ===================================================
                      ===================================================

                     ****************************************
         ************************************************************
                     ****************************************

                             Installation Details

All machine-dependent aspects of the suite have been isolated in the 
subroutine ENVIRN found in the rksuite.for file.  Some environmental
parameters must be specified in this subroutine.  The values in the
distribution version are those appropriate to the IEEE arithmetic 
standard.  They must be altered, if necessary, to values appropriate to 
the computing system you are using before calling the codes of the suite.
If the IEEE arithmetic standard values are not appropriate for your
system, appropriate values can be often be obtained by calling routines
named in the Comments of ENVIRN.

         ================================================================
         ================================================================
         TO MAKE SURE THAT YOU SPECIFY THESE MACHINE AND INSTALLATION 
         DEPENDENT QUANTITIES PROPERLY, WHEN THE DISTRIBUTION VERSION IS
         CALLED IT WRITES A MESSAGE ABOUT THE QUANTITIES TO THE STANDARD 
         OUTPUT CHANNEL THEN TERMINATES THE RUN.  THE VALUES PROVIDED IN 
         THE DISTRIBUTION VERSION SHOULD BE ALTERED, IF NECESSARY, THEN
         THE "WRITE" AND "STOP" STATEMENTS MUST BE COMMENTED OUT.
         ================================================================
         ================================================================

The distribution version of rksuite.for is in DOUBLE PRECISION. A REAL 
version is also available.  When solving ordinary differential equations 
on many popular computers, it is advisable to use DOUBLE PRECISION. However,
if DOUBLE PRECISION provides more than about 20 significant figures, the 
REAL version will usually be satisfactory, provided that the accuracy 
required of the solution is meaningful in the REAL machine precision. 

The values of the coefficients of the Runge-Kutta methods are supplied in
the subroutine CONST as DOUBLE PRECISION constants. Some of the coefficients
are given to 30 significant figures.  It is likely that your compiler will 
round these constants correctly to the nearest representable machine number.
If possible, you should request your compiler to round the constants rather
than truncate them.  Your compiler might warn you that it has shortened
the representation of the constants. The warning does not imply anything is
wrong, but you might wish to take action to avoid receiving such messages
every time you compile RKSUITE. 

                     ****************************************
         ************************************************************











------m na.brankin@na-net.ornl.gov Wed Jun 17 22:50:45 EDT 1992
From: Eric Grosse 908-582-5828 <ehg@research.att.com>
To-: na.brankin@na-net.ornl.gov
Date: Wed Jun 17 22:47:29 EDT 1992
Subject: netlib/rksuite

Richard,

Thank you very much for making this package available.  It is now installed
in netlib here, with only small modifications to filenames to fit our conventions.
An update was recently sent to the other netlib sites, so unfortunately it will
be a month or so until this propagates everywhere else.  On the other hand,
that does give you time to complain if my reorganization is unsuitable!
To see what I did,
   mail netlib@research.att.com
   send index for rksuite

Best wishes, Eric

------m na.brankin@na-net.ornl.gov Wed Jun 17 23:25:52 EDT 1992
From: Eric Grosse 908-582-5828 <ehg@research.att.com>
To-: na.brankin@na-net.ornl.gov
Date: Wed Jun 17 23:23:29 EDT 1992
Subject: netlib/ode/rksuite

After reflection, it seemed better to make this a subdirectory of the
existing ode directory, so more people could find it (and to combat
the sprawl of netlib at the top level).

So, the syntax is now   "send index for ode/rksuite"   though for ease
of use   "send index for rksuite"   will continue to work.

