CCL Home Page
Up Directory CCL note
---------- note that address of Mike Whitbeck changed. It is now:
whitbeck1@llnl.gov or whitbeck1@popcorn.llnl.gov

From whitbeck@sanjuan.wrc.unr.edu Tue May 14 15:49:40 1991
Received: from equinox.unr.edu by oscsunb.ccl.net (5.64+IDA+kva.1/901126.13)
	id AA08361; Tue, 14 May 91 15:49:26 -0400
Received: from wheeler.wrc.unr.edu (wrc.unr.edu) by equinox.unr.edu (4.1/1.34)
	id AA15201; Tue, 14 May 91 12:45:38 PDT
Received: from sanjuan.wrc.unr.edu 
	by wheeler.wrc.unr.edu (4.1//06-07-88)
	id AA14047; Tue, 14 May 91 12:45:16 PDT
Received: by sanjuan.wrc.unr.edu (4.1/SMI-4.1)
	id AA03303; Tue, 14 May 91 12:45:11 PDT
Date: Tue, 14 May 91 12:45:11 PDT
From: whitbeck@sanjuan.wrc.unr.edu (Mike Whitbeck)
Message-Id: <9105141945.AA03303@sanjuan.wrc.unr.edu>
To: aslam@sp63.csrd.uiuc.edu, bernie@ecr.mu.oz.au, dkent@ucsd.edu,
        eczam@clan.nottingham.ac.uk, edwards@cae.wisc.edu, emv@poe.aa.ox.com,
        huang@deakin.oz.au, jkl@ccl.net, jpensar@ra.abo.fi,
        kenr@peabody.iusb.indiana.edu, pc@kjemi.unit.no, peri@eng.buffalo.edu,
        probert%cs@hub.ucsb.edu, shankar@s1.msi.umn.edu
Subject: ftp site
Status: R

An ftp site for REACT (chemical kinetics simulation) has been found at 
Ohio State. As soon as I get confirmation on its final resting
place I'll let you know.

In the meanwhile look for it at
www.ccl.net (128.146.36.48)
in pub/incoming/react.tar.Z

get it in binary mode
uncompress it and untar it
zcat react.tar.Z | tar xvf -

It should make a directory called REACT.

If you are on a sun 3|4
cd REACT
make libq.a
cd reaction
make react select extract

NOTE: if you do not support -fswitch then change this in the
makefiles and use whatever hard math you do have or just leave it
off.


If you are on a VAX, well errr what can I say? If you get it ported
please let me know!

If you are using a MSDOS, ATARI, AMIGA... the port should be fairly
straight forward. Look in rxn.h for parameters that address the
size of the systems that can be handled- reduce accordingly if your
machine is shy on memory. The string functions are the most likely
place for incompatibility between unix and these other systems-
be especially attentative to fgets() and strtok() and mytok() [my
copy of strtok()].  Let me know if you port it and I'll incorporate
the changes with conditional compiler statements.



MANY THANKS to Jan K. Labanowski, Ph.D., Senior Supercomputer
Specialist for making an ftp site available.

Jan will make an announcement to a computational chemistry list at
the Ohio Supercomputer Center after he  has checked it out. At that
time it may/probably be moved to another directory.


IN EUROPE: try
(ugle.unit.no 129.241.1.97)  incoming/react.tar.Z
MANY THANKS here to Per Christer Lund              
    Department of Chemical Engineering   
  The Norwegian Institute of Technology  


Enjoy!

___________________________________________________________
|Mike Whitbeck             | whitbeck@unssun.unr.edu      |
|Desert Research Inst.     | whitbeck@wheeler.wrc.unr.edu |
|POB 60220                 | whitbeck@sanjuan.UUCP        |
|RENO, NV 89506            | 702-673-7348                 |
|__________________________|______________________________|

All academics have the potential for being insatiable...
but the chemists are the most expensive and insatiable among the
expensive and insatiable. - J. Martin in "To Rise Above Principle"


From whitbeck@sanjuan.wrc.unr.edu Tue May 14 15:59:28 1991
Received: from equinox.unr.edu by oscsunb.ccl.net (5.64+IDA+kva.1/901126.13)
	id AA08502; Tue, 14 May 91 15:59:04 -0400
Received: from wheeler.wrc.unr.edu (wrc.unr.edu) by equinox.unr.edu (4.1/1.34)
	id AA15856; Tue, 14 May 91 12:56:04 PDT
Received: from sanjuan.wrc.unr.edu 
	by wheeler.wrc.unr.edu (4.1//06-07-88)
	id AA14109; Tue, 14 May 91 12:55:41 PDT
Received: by sanjuan.wrc.unr.edu (4.1/SMI-4.1)
	id AA03369; Tue, 14 May 91 12:55:37 PDT
Date: Tue, 14 May 91 12:55:37 PDT
From: whitbeck@sanjuan.wrc.unr.edu (Mike Whitbeck)
Message-Id: <9105141955.AA03369@sanjuan.wrc.unr.edu>
To: aslam@sp63.csrd.uiuc.edu, bernie@ecr.mu.oz.au, dkent@ucsd.edu,
        eczam@clan.nottingham.ac.uk, edwards@cae.wisc.edu, emv@poe.aa.ox.com,
        huang@deakin.oz.au, jkl@ccl.net, jpensar@ra.abo.fi,
        kenr@peabody.iusb.indiana.edu, pc@kjemi.unit.no, peri@eng.buffalo.edu,
        probert%cs@hub.ucsb.edu, shankar@s1.msi.umn.edu
Subject: disclaimer
Status: R

Not being a lawyer I always forget these things.

The software I am providing comes free and with no gurantees.
Use at your own risk.

I hope to here from some of you after you have had a chance to try
it out. If I get feedback on bugs/enhancements I will try to keep
the code current and notify users of such changes.

If at anytime you want to be removed from this list just send mail.

Mike Whitbeck, Desert Research Institute.


From whitbeck@sanjuan.wrc.unr.edu Tue May 14 18:44:34 1991
Received: from equinox.unr.edu by oscsunb.ccl.net (5.64+IDA+kva.1/901126.13)
	id AA10869; Tue, 14 May 91 18:44:12 -0400
Received: from wheeler.wrc.unr.edu (wrc.unr.edu) by equinox.unr.edu (4.1/1.34)
	id AA26534; Tue, 14 May 91 15:41:10 PDT
Received: from sanjuan.wrc.unr.edu 
	by wheeler.wrc.unr.edu (4.1//06-07-88)
	id AA15324; Tue, 14 May 91 15:40:44 PDT
Received: by sanjuan.wrc.unr.edu (4.1/SMI-4.1)
	id AA03528; Tue, 14 May 91 15:40:40 PDT
Date: Tue, 14 May 91 15:40:40 PDT
From: whitbeck@sanjuan.wrc.unr.edu (Mike Whitbeck)
Message-Id: <9105142240.AA03528@sanjuan.wrc.unr.edu>
To: aslam@sp63.csrd.uiuc.edu, bernie@ecr.mu.oz.au, dkent@ucsd.edu,
        eczam@clan.nottingham.ac.uk, edwards@cae.wisc.edu, emv@poe.aa.ox.com,
        huang@deakin.oz.au, jkl@ccl.net, jpensar@ra.abo.fi,
        kenr@peabody.iusb.indiana.edu, pc@kjemi.unit.no, peri@eng.buffalo.edu,
        probert%cs@hub.ucsb.edu, rcruz@paul.rutgers.edu,
        shankar@s1.msi.umn.edu
Subject: some test data
Status: R

Here is a test problem a little larger than the example given in the
distribution for REACT.
============================= problem.m===================================
 .2500e-05	no2m		->	no + om
 .5556e+10	om		->	oh + ohm
 .1000e+10	oh + ohm	->	om
 .1000e+10	oh + no2m	->	no2 + ohm
 .2500e+09	om + no2m	->	no2 + ohm + ohm
 .2500e+10	om + o2		->	o3m
 .5500e+04	o3m 		->	o2 + om
 .1000e+09	no2 + no2	->	n2o4
 .1300e+04	n2o4		->	no2 + no2
 .1200e+04	n2o4		->	no2m + no3m
 .3000e+03	no2		->	x
 .1000e+09	no2 + x		->	no2m + no3m
 .1000e+09	no2 + no	->	n2o3
 .2000e+04	n2o3		->	no2 + no
 .2000e+04	n2o3		->	no2m + no2m
 .1000e-00	x + no		->	no2m + no2m
============================= problem.m===================================
and here are the parameters.
============================= problem.p===================================
0 .36e+04 .06e+04 1.0e-03
no2m	0.1e-0	1.0e-8	1.0e-03
o2	.3e-03	1.0e-8	1.0e-03
ohm	.1e-06	1.0e-8 1.0e-03
============================= problem.p===================================
The solution is found by
...............................................................
%react problem
1)  .2500e-05	no2m		->	no + om
2)  .5556e+10	om		->	oh + ohm
3)  .1000e+10	oh + ohm	->	om
4)  .1000e+10	oh + no2m	->	no2 + ohm
5)  .2500e+09	om + no2m	->	no2 + ohm + ohm
6)  .2500e+10	om + o2		->	o3m
7)  .5500e+04	o3m 		->	o2 + om
8)  .1000e+09	no2 + no2	->	n2o4
9)  .1300e+04	n2o4		->	no2 + no2
10)  .1200e+04	n2o4		->	no2m + no3m
11)  .3000e+03	no2		->	x
12)  .1000e+09	no2 + x		->	no2m + no3m
13)  .1000e+09	no2 + no	->	n2o3
14)  .2000e+04	n2o3		->	no2 + no
15)  .2000e+04	n2o3		->	no2m + no2m
16)  .1000e-00	x + no		->	no2m + no2m
...............................................................
6.340u 0.340s 0:12.26 54% 0+22k 0+2io 0pf+0w
(6.34 seconds run time)

I can look at the results using select (here I have a file problem.e
with only the words 
no2 no no3m
rather than everything to reduce the amount of output)
...............................................................
% select no2aq
16 reacions involving 12 species
no2m  no  om  oh  ohm  no2  o2  o3m  n2o4  no3m  x  n2o3 
       time            no2             no           no3m
     0.0000e+00      0.0000e+00      0.0000e+00      0.0000e+00 
     6.0000e+02      8.5578e-11      4.5160e-05      2.1075e-05 
     1.2000e+03      6.7257e-11      6.5957e-05      3.1552e-05 
     1.8000e+03      5.2981e-11      8.3997e-05      4.0644e-05 
     2.4000e+03      4.6293e-11      9.9261e-05      4.8192e-05 
     3.0000e+03      4.1795e-11      1.1373e-04      5.5295e-05 
     3.6000e+03      3.8639e-11      1.2502e-04      6.0946e-05 
...............................................................

Now how do I know that the tolerance values are ok?  I could rerun
with smaller values and look for an effect. But in this case I was
carefull to write reactions conserving all 'n' containing species (not
so with o an h, in this sample I make OH- readily (ohm) but not H+
where I should!). If I add up all the 'n's in the result I should get
what I started with (0.10 == no2m initial, see problem.p).  Here is
the problem.e file to extract all the species with 'n's:

=============================problem.e===================================
 no2m  no  no2  n2o4 n2o4  no3m  x  n2o3 n2o3
=============================problem.e===================================

notice n2o4 and n2o3 are in the list twice so they get counted twice!
I just pipe the output to Gary Pearlman's dm to add them up!
(recall that the first column is time)

%extract problem | dm ' x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9 + x10 '

to get:
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1

You could also use an awk script

% extract problem | awk '{a= $2+$3+$4+$5+$6+$7+$8+$9+$10; print a}'
.1
.1
.1
.1
.1
.1
.1
...............................................................


You may have noticed that since I use '-' as a token in the reaction
arrow I designate charge with a lowercase m or p appended to the name.

Tokens parsed as word delimeters in the mechanism file are:
	"$,+;:(){}[]"  and "->"

That is 
	[CH3] + (CH2) -> {stuff}
is equivalent to
	CH3 + CH2 -> stuff

In a future release I may remove the "(){}[]" from the token list after
I decide what if anything I want to do with automated invocation of
the steady state approximation (I was thinking about enclosing
variables in {} if they could be approximated by ss).

___________________________________________________________
|Mike Whitbeck             | whitbeck@unssun.unr.edu      |
|Desert Research Inst.     | whitbeck@wheeler.wrc.unr.edu |
|POB 60220                 | whitbeck@sanjuan.UUCP        |
|RENO, NV 89506            | 702-673-7348                 |
|__________________________|______________________________|

All academics have the potential for being insatiable...
but the chemists are the most expensive and insatiable among the
expensive and insatiable. - J. Martin in "To Rise Above Principle"

From whitbeck@sanjuan.wrc.unr.edu Wed May 15 14:57:18 1991
Received: from equinox.unr.edu by oscsunb.ccl.net (5.64+IDA+kva.1/901126.13)
	id AA07543; Wed, 15 May 91 14:56:54 -0400
Received: from wheeler.wrc.unr.edu (wrc.unr.edu) by equinox.unr.edu (4.1/1.34)
	id AA03211; Wed, 15 May 91 11:46:47 PDT
Received: from sanjuan.wrc.unr.edu 
	by wheeler.wrc.unr.edu (4.1//06-07-88)
	id AA18235; Wed, 15 May 91 11:46:22 PDT
Received: by sanjuan.wrc.unr.edu (4.1/SMI-4.1)
	id AA03993; Wed, 15 May 91 11:46:15 PDT
Date: Wed, 15 May 91 11:46:15 PDT
From: whitbeck@sanjuan.wrc.unr.edu (Mike Whitbeck)
Message-Id: <9105151846.AA03993@sanjuan.wrc.unr.edu>
To: aslam@sp63.csrd.uiuc.edu, beasockj@lonex.radc.af.mil, bernie@ecr.mu.oz.au,
        dkent@ucsd.edu, eczam@clan.nottingham.ac.uk, edwards@cae.wisc.edu,
        emv@poe.aa.ox.com, huang@deakin.oz.au, jkl@ccl.net, jpensar@ra.abo.fi,
        kenr@peabody.iusb.indiana.edu, klemm@msc.edu, mderksen@sci.kun.nl,
        nanis@llex.ll.mit.edu, pc@kjemi.unit.no, peri@eng.buffalo.edu,
        probert%cs@hub.ucsb.edu, rcruz@paul.rutgers.edu,
        shankar@s1.msi.umn.edu
Subject: my error !
Status: R

The response to the REACT software has been overwhelming! If I get
a little behind in answering individual mail please have patience;
I'll get to it! Here are some frequently asked questions:

1) fortran version
There was ONCE a fortran version that was a very complete tool for
the analysis of chemical kinetic problems. It is NO LONGER
AVAILABLE (I lost it!). The computer center is checking an archive
tape but there is little hope. I have a VERY LONG listing if there
is interest I can make copies available. As time goes by I hope to
add all the features that were in the fortran code to REACT.

2) example problem works, extract works but select fails

To my chagrin I goofed in the test problem recently mailed.

After running the problem I said enter
	select no2aq
to view some results... but I told you to name the test files
'problem' so that should be
	select problem

select then looks for problem.e for a list of species to tabulate
vs time. If you entered select no2aq it looked for and failed to
find no2aq.e for such a list.

3) port to DOS and DOS like critters

I have started the port to the ATARI 1040ST. In the process I
discovered a problem. I guess I am too use to the SUN on my desk
because I have become lazy about checking for free memory. There 
are a lot of malloc's in the code so one of the next revisions will
test each malloc for success/failure. This will help those who want
to port to PC's or other memory stingy platforms.


Programmers note: If you want you can replace the call to lsoda
with your own integrater such as a Runge-Kutta routine. It will
probably run much slower but will be much smaller code.

==========
I'll try not to bother you with many more of these posting's. 
Please let me know if you have any problems etc.

Thanks for the interest.

___________________________________________________________
|Mike Whitbeck             | whitbeck@unssun.unr.edu      |
|Desert Research Inst.     | whitbeck@wheeler.wrc.unr.edu |
|POB 60220                 | whitbeck@sanjuan.UUCP        |
|RENO, NV 89506            | 702-673-7348                 |
|__________________________|______________________________|

All academics have the potential for being insatiable...
but the chemists are the most expensive and insatiable among the
expensive and insatiable. - J. Martin in "To Rise Above Principle"


From whitbeck@sanjuan.wrc.unr.edu Fri May 17 20:53:10 1991
Received: from equinox.unr.edu by oscsunb.ccl.net (5.64+IDA+kva.1/901126.13)
	id AA23015; Fri, 17 May 91 20:52:11 -0400
Received: from wheeler.wrc.unr.edu (wrc.unr.edu) by equinox.unr.edu (4.1/1.34)
	id AA19670; Fri, 17 May 91 15:43:39 PDT
Received: from sanjuan.wrc.unr.edu 
	by wheeler.wrc.unr.edu (4.1//06-07-88)
	id AA24945; Fri, 17 May 91 15:43:14 PDT
Received: by sanjuan.wrc.unr.edu (4.1/SMI-4.1)
	id AA05216; Fri, 17 May 91 15:43:09 PDT
Date: Fri, 17 May 91 15:43:09 PDT
From: whitbeck@sanjuan.wrc.unr.edu (Mike Whitbeck)
Message-Id: <9105172243.AA05216@sanjuan.wrc.unr.edu>
To: TTENSFELDT@eagle.wesleyan.edu, aslam@sp63.csrd.uiuc.edu,
        beasockj@lonex.radc.af.mil, bernie@ecr.mu.oz.au, dkent@ucsd.edu,
        dove@cats.ucsc.edu, eczam@clan.nottingham.ac.uk, edwards@cae.wisc.edu,
        emv@poe.aa.ox.com, huang@deakin.oz.au, jkl@ccl.net, jpensar@ra.abo.fi,
        kenr@peabody.iusb.indiana.edu, klemm@msc.edu, mderksen@sci.kun.nl,
        mgo@inel.gov, mr0183@usma2.usma.edu, nanis@llex.ll.mit.edu,
        pc@kjemi.unit.no, peri@eng.buffalo.edu, probert%cs@hub.ucsb.edu,
        rcruz@paul.rutgers.edu, shankar@s1.msi.umn.edu, water@u.washington.edu,
        yos@tasman.cc.utas.edu.au
Subject: lsoda ... more than 500 steps
Status: R



IF in using REACT you should see a message like

lsoda .... more than 500 steps taken....tout ...


this is a complaint from the core integrater that it is having a
difficult time achieving the desired accuracy.

In  a future release I will trap on this condition to stop the
program. In my experience thus far this condition does not
easily cure itself. For now I suggest that when you see this (or
any message from lsoda) you kill the job. Take a look at the 
tolerance values that you specified. Typically they need to be
numerically larger (lower tolerance) to correct this problem.

===

some of you have noticed that in the distribution there is a
program called example.c in with the lsoda stuff. I hacked this
from netlibs example.f for testing lsoda. If you want to be
impressed by lsoda's abilities compile example.c and link in
the library made with lsoda
	cc -o wow example.c -lq -lm -fswitch
	wow
[leave off -fswitch if not supported at your site]
The first column in the output is the independent variable
(time); note the time steps!


Modified: Tue Mar 28 17:00:00 1995 GMT
Page accessed 5207 times since Sat Apr 17 21:32:47 1999 GMT