File: README

package info (click to toggle)
pgapack 1.0.0.1-2.1
  • links: PTS
  • area: main
  • in suites: woody
  • size: 2,348 kB
  • ctags: 1,780
  • sloc: ansic: 10,331; fortran: 2,985; sh: 574; makefile: 431; perl: 105
file content (231 lines) | stat: -rw-r--r-- 9,893 bytes parent folder | download | duplicates (5)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
This is the top-level README for PGAPack V1.0

Copyright
=========
See the file COPYRIGHT for Copyright and disclaimer information.

Introduction
============
PGAPack is a general-purpose, data-structure-neutral, parallel genetic
algorithm library being developed at Argonne National Laboratory.  
Key features are:

* Callable from Fortran or C.
* Runs on uniprocessors, parallel computers, and workstation networks.
* Binary-, integer-, real-, and character-valued native data types.
* Object-oriented data structure neutral design. 
* Parameterized population replacement.
* Multiple choices for selection, crossover, and mutation operators.
* Easy integration of hill-climbing heuristics. 
* Easy-to-use interface for novice and application users.
* Fully extensible to support custom operators and new data types.
* Extensive debugging facilities.
* A large set of example problems.


Availability
============
PGAPack is freely available.  The complete distribution is available
by anonymous ftp from from ftp.mcs.anl.gov in the file 
pub/pgapack/pgapack.tar.Z.  The distribution contains all source code,
installation instructions, users guide, and a collection of examples in C and
Fortran. Additional information is available on the World Wide Web at
http://www.mcs.anl.gov/pgapack.html.


Computational Environment
=========================
PGAPack is written in ANSI C and uses the MPI message passing interface
and should run on most uniprocessors, parallel computers, and workstation
networks.  PGAPack has been tested on the workstations and parallel computers 
specified by the ARCH_TYPE variable below.


Documentation
=============
* The PGAPack users guide is in ./docs/user_guide.ps.
* Man pages for PGAPack functions are in the ./man directory.
* Installation instructions are in this README file and the users guide.
* Example problems are in the ./examples directory.


Installation Requirements
=========================
To compile you must have an ANSI C compiler that includes a full
implementation of the Standard C library and related header files.  To build a
*parallel* version of PGAPack you must provide an implementation of MPI
(Message Passing Interface) for the parallel computer or workstation network
you are running on.

Most of our testing and development was done using MPICH, a freely available
implementation of MPI.  MPICH runs on many parallel computers and
workstation networks and is publicly available and free.  The complete
distribution is available by anonymous ftp from ftp.mcs.anl.gov.  Take the
file mpich.tar.Z from the directory pub/mpi.  Additional information about
MPICH is avaliable on the World Wide Web at http://www.mcs.anl.gov/mpi.

Installation Instructions
=========================
When installing PGAPack you make two choices: whether to build a sequential
(the default) or parallel version, and whether to build a debug or optimized
(the default) version.  In broad outline, the
installation steps are as follows.

1.  Make a directory to install PGAPack in (mkdir /usr/local/pga).
2.  Change directories to the directory created in the last step
    (cd /usr/local/pga).
3.  Obtain the compressed tar file pgapack.tar.Z by anonymous ftp in
    ftp.mcs.anl.gov from the directory pub/pgapack.
4.  Uncompress the tar file (uncompress pgapack.tar.Z).
5.  Untar the uncompressed tar file (tar xvf ./pgapack.tar).
6.  Run configure to configure the makefiles (configure -arch ARCH_TYPE)
    where ARCH_TYPE is one of

    sun4           for Sun SparcStations workstations,
    next           for NeXT workstations,
    rs600          for IBM RS6000 workstations,
    irix           for Silicon Graphics workstations,
    hpux           for Hewlett Packard workstations,
    alpha          for DEC Alpha workstations,
    linux          for machines running Linux,
    freebsd        for machines running FreeBSD,
    generic        for generic 32-bit machines, 
    powerchallenge for the Silicon Graphics Power Challenge Array,
    challenge      for the Silicon Graphics Challenge,
    t3d            for the Cray T3D,
    sp2            for the IBM SP2,
    paragon        for the Intel Paragon, or
    exemplar       for the Convex  Exemplar.

    The full configure options are configure -arch ARCH_TYPE [-cc CC] 
    [-cflags CFLAGS] [-f77 FC] [-fflags FFLAGS] [-debug]
    [-mpiinc MPI_INCLUDE_DIRECTORY] [-mpilib MPI_LIBRARY]  [-help]

    where all parameters except -arch are optional and do the following:

    [-cc]:       The name of the ANSI C compiler, cc by default.
    [-cflags]:   Options passed to the C compiler.
    [-f77]:      The name of the Fortran 77 compiler, f77 by default.
                 (The Fortran compiler is used only to compile the Fortran
                 examples in the [./examples/] directory.)
    [-fflags]:   Options passed to the Fortran compiler.
    [-debug]:    If specified, enables the debugging features
                 and compiles the source code with the -g flag.
    [-mpiinc]:   The *directory* where MPI include files are located.
    [-mpilib]:   The *full path* to the MPI library.  

    If the [-mpilib] and [-mpiinc] options are specified, a parallel version 
    of PGAPack will be built.  If these flags are not specified a sequential
    version of PGAPack will be built.

7.  Execute the makefile (make install).
8.  Add PGAPack's man pages to your man page path.
    (setenv MANPATH "$MANPATH"":/home/pgapack/man")
9.  Execute a simple test problem
    
    Sequential version
    ------------------
    C:        /usr/local/pga/examples/c/maxbit
    Fortran:  /usr/local/pga/examples/fortran/maxbit

    Parallel version
    ------------------
    C:        mpirun -np 4 /usr/local/pga/examples/c/maxbit
    Fortran:  mpirun -np 4 /usr/local/pga/examples/fortran/maxbit

    If a parallel version of PGAPack was used, the actual commands to execute 
    a parallel program depend on the particular MPI implementation and
    parallel computer.  If the MPICH implementation was used the "mpirun"
    command can be used to execute a parallel program on most systems.


Example Installation
====================
Assuming you have the compressed tar file, pgapack.tar.Z, containing the
distribution in /home/username.  To build a sequential version of PGAPack
containing debug features for a Sun SparcStation in /usr/local/pga and run a
test example, type:

    uncompress /home/username/pgapack.tar.Z
    mkdir /usr/local/pga
    cd /usr/local/pga
    tar xvf /home/username/pgapack.tar
    configure -arch sun4 -debug
    make install
    /usr/local/pga/examples/c/maxbit

To build a parallel version of PGAPack for an IBM SP parallel computer using
an MPI implementation with include files in /usr/local/mpi/include and library
in /usr/local/mpi/lib, and run a test example, type:

   uncompress /home/username/pgapack.tar.Z
   mkdir /usr/local/pga
   cd /usr/local/pga
   tar xvf /home/username/pgapack.tar
   configure -arch rs6000 -mpiinc /usr/local/mpi/include \
             -mpilib /usr/local/mpi/lib/libmpi.a
   make install
   mpirun -np 4 /usr/local/pga/examples/c/maxbit


PGAPack on PCs 
===================
PGAPack has not been ported to MS-DOS, Windows 3.1, Windows 95, or Apple OS.
As mentioned earlier, however, PGAPack is written in ANSI standard C and
should compile in these environments.  Be aware, however, that PGAPack's
random number generator, PGARandom01, assumes certain machine characteristics
for ints and floats that may not correspond with what your PC and/or compiler
support, resulting in erroneous values.


Structure of the Distribution Directory
=======================================
CHANGES:      Changes new to this release of PGAPack.
COPYRIGHT:    Copyright and disclaimer information.
README:       This file.
configure.in: The ``source code'' for the configure script.
configure:    Unix shell script to configure Makefile.in for a specific
              architecture.
Makefile.in:  Prototype that is configured into ``Makefile'' by configure.
docs:         Directory containing documentation.
examples:     A directory containing C and Fortran examples.
include:      The PGAPack include directory.
lib:          The directory the library will be installed in.
man:          The directory containing the PGAPack man pages.
source:       The source code for the PGAPack system.
test:         A directory containing programs to verify the installation.


Email Addresses
===============
To join the PGAPack mailing list to receive announcements of new versions,
enhancements, and bug fixes, send electronic mail to pgapack@mcs.anl.gov.  Bug
reports should be sent to pgapack-bugs@mcs.anl.gov.  The World Wide Web page
for PGAPack is http://www.mcs.anl.gov/pgapack.html and contains up-to-date
news and a list of outstanding bug reports.


When reporting a bug please include as much information and documentation as
possible.  Helpful information would include PGAPack version number, MPI
implementation and version used (if any), configuration options, type of
computer system, problem description and error message output.  It is helpful
if you put a PGAPrintContextVariable call before and after the PGASetUp call.
Additionally, if possible, build a debug version of PGAPack and send
``high-level'' output from running your program with the trace facility
enabled.


Contributions
=============
PGAPack was written to be extensible in two ways: adding new operators that
work with existing data types, and defining new data types.  Enhancements of
either type that you wish to share are welcome for possible inclusion in
future versions of PGAPack.


Acknowledgment
==============
Users of PGAPack are asked to acknowledge its use in any document referencing
work based on the program, such as published research.  Also, please supply
to us a copy of any published research referencing work based on the software.