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
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
|
\par
\section{Driver programs for the {\tt GPart} object}
\label{section:GPart:drivers}
\par
This section contains brief descriptions of four driver programs.
\par
%=======================================================================
\begin{enumerate}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testDDviaFishnet msglvl msgFile inGraphFile freeze minweight maxweight
seed outIVfile
\end{verbatim}
This driver program constructs a domain decomposition via the {\it
fishnet} algorithm \cite{ash97-DDSEP}.
It reads in a {\tt Graph} object from a file, finds the domain
decomposition using the four input parameters, then optionally
writes out the map from vertices to components to a file.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
\item
The {\tt msgFile} parameter determines the output file --- if {\tt
msgFile} is {\tt stdout}, then the output file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inGraphFile} parameter is the input file for the {\tt Graph}
object. It must be of the form {\tt *.graphf} or {\tt *.graphb}.
The {\tt Graph} object is read from the file via the
{\tt Graph\_readFromFile()} method.
\item
The {\tt freeze} parameter is used to place nodes of high degree
into the multisector.
If the external degree of a vertex is {\tt freeze} times the
average degree, then it is placed in the multisector.
\item
The {\it target} minimum weight for a domain is {\tt minweight}.
\item
The {\it target} maximum weight for a domain is {\tt maxweight}.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt outIVfile} parameter is the output file for the {\tt IV}
object that contains the map from vertices to components.
If {\tt outIVfile} is {\tt "none"}, then there is no output,
otherwise {\tt outIVfile}
must be of the form {\tt *.ivf} or {\tt *.ivb}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testTwoSetViaBKL msglvl msgFile inGraphFile inIVfile
seed alpha outIVfile
\end{verbatim}
This driver program constructs a two-set partition via the
Block Kernighan-Lin algorithm \cite{ash97-DDSEP}.
It reads in a {\tt Graph} object and an {\tt IV} object that holds
the map from vertices to components (e.g., the output from the
driver program {\tt testDDviaFishet}) from two files,
constructs the domain-segment graph and finds an initial
separator, then optionally
writes out the new map from vertices to components to a file.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
\item
The {\tt msgFile} parameter determines the output file --- if {\tt
msgFile} is {\tt stdout}, then the output file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inGraphFile} parameter is the input file for the {\tt Graph}
object. It must be of the form {\tt *.graphf} or {\tt *.graphb}.
The {\tt Graph} object is read from the file via the
{\tt Graph\_readFromFile()} method.
\item
The {\tt inIVfile} parameter is the input file for the {\tt IV}
object that contains the map from vertices to domains and multisector.
It {\tt inIVfile} must be of the form {\tt *.ivf} or {\tt *.ivb}.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt alpha} parameter controls the partition evaluation function.
\item
The {\tt outIVfile} parameter is the output file for the {\tt IV}
object that contains the map from vertices to separator and the
two components.
If {\tt outIVfile} is {\tt "none"}, then there is no output,
otherwise {\tt outIVfile}
must be of the form {\tt *.ivf} or {\tt *.ivb}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testSmoothBisector msglvl msgFile inGraphFile inIVfile
option alpha outIVfile
\end{verbatim}
This driver program smooths a bisector of a graph by solving a
sequence of max-flow network problems.
It reads in a {\tt Graph} object and an {\tt IV} object that holds
the map from vertices to components (e.g., the output from the
driver program {\tt testTwoSetViaBKL}) from two files,
smooths the separator and
then optionally writes out the new component ids map to a file.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
\item
The {\tt msgFile} parameter determines the output file --- if {\tt
msgFile} is {\tt stdout}, then the output file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inGraphFile} parameter is the input file for the {\tt Graph}
object. It must be of the form {\tt *.graphf} or {\tt *.graphb}.
The {\tt Graph} object is read from the file via the
{\tt Graph\_readFromFile()} method.
\item
The {\tt inIVfile} parameter is the input file for the {\tt IV}
object that contains the map from vertices to domains and multisector.
It {\tt inIVfile} must be of the form {\tt *.ivf} or {\tt *.ivb}.
\item
The {\tt option} parameter specifies the type of network
optimization problem that will be solved.
\begin{itemize}
\item
{\tt option = 1} --- each network has two layers and is bipartite.
\item
{\tt option = 2}
--- each network has two layers but need not be bipartite.
\item
{\tt option = 2} --- each network has {\tt option/2} layers
on each side of the separator.
\end{itemize}
\item
The {\tt alpha} parameter controls the partition evaluation function.
\item
The {\tt outIVfile} parameter is the output file for the {\tt IV}
object that contains the map from vertices to separator and the
two components.
If {\tt outIVfile} is {\tt "none"}, then there is no output,
otherwise {\tt outIVfile}
must be of the form {\tt *.ivf} or {\tt *.ivb}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testRBviaDDsep msglvl msgFile inGraphFile seed minweight maxweight
freeze alpha maxdomweight DDoption nlayer
testRBviaDDsep2 msglvl msgFile inGraphFile nruns seed minweight maxweight
freeze alpha maxdomweight DDoption nlayer
\end{verbatim}
These driver programs construct a multisector via recursive
bisection and orders the graph using nested dissection and
multisection using the multisector.
{\tt testRBviaDDsep} executes only one run while
{\tt testRBviaDDsep2} executes {\tt nruns} runs with random
permutations of the graph.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
\item
The {\tt msgFile} parameter determines the output file --- if {\tt
msgFile} is {\tt stdout}, then the output file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inGraphFile} parameter is the input file for the {\tt Graph}
object. It must be of the form {\tt *.graphf} or {\tt *.graphb}.
The {\tt Graph} object is read from the file via the
{\tt Graph\_readFromFile()} method.
\item
The {\tt nruns} parameter is the number of runs made with the
graph randomly permuted.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\it target} minimum weight for a domain is {\tt minweight}.
\item
The {\it target} maximum weight for a domain is {\tt maxweight}.
\item
The {\tt freeze} parameter is used to place nodes of high degree
into the multisector.
If the external degree of a vertex is {\tt freeze} times the
average degree, then it is placed in the multisector.
\item
The {\tt alpha} parameter controls the partition evaluation function.
\item
The {\tt maxdomweight} parameter controls the recursive bisection
--- no subgraph with weight less than {\tt maxdomweight} is
further split.
\item
The {\tt DDoption} parameter controls the initial domain/segment
partition on each subgraph.
When {\tt DDDoption = 1} we use the fishnet algorithm for each
subgraph.
When {\tt DDDoption = 1} we use the fishnet algorithm once for the
entire graph and this is then projected down onto each subgraph.
\item
The {\tt nlayer} parameter governs the smoothing process by specifying
the type of network optimization problem that will be solved.
\begin{itemize}
\item
{\tt nlayer = 1} --- each network has two layers and is bipartite.
\item
{\tt nlayer = 2}
--- each network has two layers but need not be bipartite.
\item
{\tt nlayer > 2} --- each network has {\tt option/2} layers
on each side of the separator.
\end{itemize}
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
mkDSTree msglvl msgFile inGraphFile seed minweight maxweight
freeze alpha maxdomweight DDoption nlayer outDSTreeFile
\end{verbatim}
This driver program constructs a domain/separator tree using
recursive bisection.
The {\tt DSTree} object is optionally written to a file.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
\item
The {\tt msgFile} parameter determines the output file --- if {\tt
msgFile} is {\tt stdout}, then the output file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inGraphFile} parameter is the input file for the {\tt Graph}
object. It must be of the form {\tt *.graphf} or {\tt *.graphb}.
The {\tt Graph} object is read from the file via the
{\tt Graph\_readFromFile()} method.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\it target} minimum weight for a domain is {\tt minweight}.
\item
The {\it target} maximum weight for a domain is {\tt maxweight}.
\item
The {\tt freeze} parameter is used to place nodes of high degree
into the multisector.
If the external degree of a vertex is {\tt freeze} times the
average degree, then it is placed in the multisector.
\item
The {\tt alpha} parameter controls the partition evaluation function.
\item
The {\tt maxdomweight} parameter controls the recursive bisection
--- no subgraph with weight less than {\tt maxdomweight} is
further split.
\item
The {\tt DDoption} parameter controls the initial domain/segment
partition on each subgraph.
When {\tt DDDoption = 1} we use the fishnet algorithm for each
subgraph.
When {\tt DDDoption = 1} we use the fishnet algorithm once for the
entire graph and this is then projected down onto each subgraph.
\item
The {\tt nlayer} parameter governs the smoothing process by specifying
the type of network optimization problem that will be solved.
\begin{itemize}
\item
{\tt nlayer = 1} --- each network has two layers and is bipartite.
\item
{\tt nlayer = 2}
--- each network has two layers but need not be bipartite.
\item
{\tt nlayer > 2} --- each network has {\tt option/2} layers
on each side of the separator.
\end{itemize}
\item
The {\tt outDSTreeFile} parameter is the output file
for the {\tt DSTree} object.
It must be of the form {\tt *.dstreef} or {\tt *.dstreeb}.
If {\tt outDSTreeFile} is not {\tt "none"},
the {\tt DSTree} object is written to the file via the
{\tt DSTree\_writeToFile()} method.
\end{itemize}
%-----------------------------------------------------------------------
\end{enumerate}
|
