File: drivers.tex

package info (click to toggle)
spooles 2.2-5
  • links: PTS
  • area: main
  • in suites: lenny
  • size: 18,824 kB
  • ctags: 3,665
  • sloc: ansic: 146,828; csh: 3,615; makefile: 2,045; perl: 70
file content (827 lines) | stat: -rw-r--r-- 33,766 bytes parent folder | download | duplicates (7)
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
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
\par
\section{Driver programs for the {\tt ETree} object}
\label{section:ETree:drivers}
\par
This section contains brief descriptions of the driver programs.
\par
%=======================================================================
\begin{enumerate}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
createETree msglvl msgFile inGraphFile inPermFile outIVfile outETreeFile
\end{verbatim}
This driver program reads in a {\tt Graph} object and a {\tt Perm}
permutation object and creates a front tree {\tt ETree} object.
The map from vertices to fronts is optionally written out to {\tt
outIVfile}.
The {\tt ETree} object is optionally written out to {\tt outETreeFile}.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message 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 inPermFile} parameter is the input file for the {\tt Perm}
object. It must be of the form {\tt *.permf} or {\tt *.permb}.
The {\tt Perm} object is read from the file via the
{\tt Perm\_readFromFile()} method.
\item
The {\tt outIVfile} parameter is the output file for the 
vertex-to-front map {\tt IV} object. 
If {\tt outIVfile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outIVfile} is of the form 
{\tt *.ivf}), or
a binary file (if {\tt outIVfile} is of the form {\tt *.ivb}).
\item
The {\tt outETreeFile} parameter is the output file for the 
{\tt ETree} object. 
If {\tt outETreeFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to 
a formatted file (if {\tt outETreeFile} is of the form 
{\tt *.etreef}),
or
a binary file (if {\tt outETreeFile} is of the form {\tt *.etreeb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
extractTopSep msglvl msgFile inETreeFile outIVfile
\end{verbatim}
This driver program creates an {\tt IV} object that contains a 
{\tt compids[]} vector, where {\tt compids[v] = 0} if vertex {\tt v} 
is in the top level separator and {\tt -1} otherwise.
The {\tt IV} object is optionally written out to a file.
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means that all objects are written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
message data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt outIVfile} parameter is the output file for the 
vertex-to-front map {\tt IV} object. 
If {\tt outIVfile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outIVfile} is of the form 
{\tt *.ivf}), or
a binary file (if {\tt outIVfile} is of the form {\tt *.ivb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
mkNDETree msglvl msgFile n1 n2 n3 maxzeros maxsize outFile
\end{verbatim}
\par
This program constructs a front tree for a Laplacian operator on a
regular grid ordered using nested dissection.
When {\tt n3 = 1}, the problem is two dimensional and a 9-point
operator is used.
When {\tt n3 > 1}, the problem is three dimensional and a 27-point
operator is used.
A sequence of five ETree objects are produced:
\begin{itemize}
\item vertex elimination tree
\item fundamental supernode front tree
\item front tree after trying to merge with an only child
\item front tree after trying to merge with all children
\item front tree after splitting large fronts
\end{itemize}
The merging and splitting process are controlled by the 
{\tt maxzeros} and {\tt maxsize} parameters.
Here is some typical output for a $15 \times 15 \times 15$ grid
matrix with {\tt maxzeros = 64} and {\tt maxsize = 32}.
\begin{verbatim}
 vtx tree :  3375 fronts,   367237 indices,   367237 |L|, 63215265 ops
 fs tree  :  1023 fronts,    39661 indices,   367237 |L|, 63215265 ops
 merge1   :  1023 fronts,    39661 indices,   367237 |L|, 63215265 ops
 merge2   :   511 fronts,    29525 indices,   373757 |L|, 63590185 ops
 split    :   536 fronts,    34484 indices,   373757 |L|, 63590185 ops
\end{verbatim}
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
{\tt n1} is the number of grid points in the first direction.
\item
{\tt n2} is the number of grid points in the second direction.
\item
{\tt n3} is the number of grid points in the third direction.
\item
The {\tt maxzeros} parameter is an upper bound on the number of
logically zero entries that will be allowed in a new front.
\item
The {\tt maxsize} parameter is an upper bound on the number of
vertices in a front --- any original front that contains more than
{\tt maxsize} vertices will be broken up into smaller fronts.
\item
The {\tt outFile} parameter is the output file for the {\tt ETree}
object. 
If {\tt outFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outFile} 
is of the form {\tt *.etreef}), or a binary file 
(if {\tt outFile} is of the form {\tt *.etreeb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
mkNDoutput msglvl msgFile n1 n2 n3 maxzeros maxsize 
           nthread maptype cutoff outETreeFile outMapFile
\end{verbatim}
\par
This program constructs a front tree for a Laplacian operator on a
regular grid ordered using nested dissection.
When {\tt n3 = 1}, the problem is two dimensional and a 9-point
operator is used.
When {\tt n3 > 1}, the problem is three dimensional and a 27-point
operator is used.
The front tree is generated in the same fashion as done by the 
{\tt mkNDETree} driver program.
Using this front tree, an {\tt IV} object that maps fronts 
to processors is then created using one of four different kinds of
maps.
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
{\tt n1} is the number of grid points in the first direction.
\item
{\tt n2} is the number of grid points in the second direction.
\item
{\tt n3} is the number of grid points in the third direction.
\item
The {\tt maxzeros} parameter is an upper bound on the number of
logically zero entries that will be allowed in a new front.
\item
The {\tt maxsize} parameter is an upper bound on the number of
vertices in a front --- any original front that contains more than
{\tt maxsize} vertices will be broken up into smaller fronts.
\item
The {\tt nthread} parameter is the number of threads.
\item
The {\tt maptype} parameter is the type of map.
\begin{itemize}
\item {\tt 1} --- wrap map
\item {\tt 2} --- balanced map
\item {\tt 3} --- subtree-subset map
\item {\tt 4} --- domain decomposition map
\end{itemize}
\item
The {\tt cutoff} parameter is used by the domain decomposition map only.
Try setting {\tt cutoff = 1/nthread} or {\tt cutoff = 1/(2*nthread)}.
\item
The {\tt outETreeFile} parameter is the output file for the {\tt ETree}
object. 
If {\tt outETreeFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outETreeFile} 
is of the form {\tt *.etreef}), or a binary file 
(if {\tt outETreeFile} is of the form {\tt *.etreeb}).
\item
The {\tt outMapFile} parameter is the output file for the {\tt IV}
map object. 
If {\tt outMapFile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outMapFile} 
is of the form {\tt *.ivf}), or a binary file 
(if {\tt outMapFile} is of the form {\tt *.ivb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
permuteETree msglvl msgFile inETreeFile inEqmapIVfile outETreeFile outIVfile
\end{verbatim}
This driver program is used to get an old-to-new permutation vector
from an {\tt ETree} object and permute the vertices in the 
{\tt ETree} object.
The program has the ability to handle an {\tt ETree} object that is
defined on a compressed graph.
If {\tt inEqmapIVfile} is not {\tt none}, the program reads in an
{\tt IV} object that contains the equivalence map, i.e., the map
from the degrees of freedom to the vertices in the compressed graph.
This map is used to expand the {\tt ETree} object.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt inEqmapIVfile} parameter is the input file for the 
equivalence map {\tt IV} object. 
It must be of the form {\tt *.ivf}, {\tt *.ivb}, or {\tt none}.
If {\tt inEqmapIVfile} is not {\tt none},
the {\tt IV} object is read from the file via the
{\tt IV\_readFromFile()} method.
\item
The {\tt outETreeFile} parameter is the output file for the {\tt ETree}
object. 
If {\tt outETreeFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outETreeFile} 
is of the form {\tt *.etreef}), or a binary file 
(if {\tt outETreeFile} is of the form {\tt *.etreeb}).
\item
The {\tt outIVFile} parameter is the output file for the 
old-to-new {\tt IV} object. 
If {\tt outIVFile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outIVFile} 
is of the form {\tt *.ivf}), or a binary file 
(if {\tt outIVFile} is of the form {\tt *.ivb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testExpand msglvl msgFile inETreeFile inEqmapFile outETreeFile
\end{verbatim}
This driver program is used to translate an {\tt ETree} object for
a compressed graph into an {\tt ETree} object for the unit weight
graph.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object for the compressed graph. 
It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt inEqmapFile} parameter contains the map from vertices 
in the unit weight graph into vertices in the compressed graph.
It must be of the form {\tt *.ivf} or {\tt *.ivb}.
The {\tt IV} object is read from the file via the
{\tt IV\_readFromFile()} method.
\item
The {\tt outETreeFile} parameter is the output file for the {\tt ETree}
object for the unit weight graph. 
If {\tt outETreeFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to 
a formatted file (if {\tt outETreeFile} is of the form {\tt *.etreef}),
or a binary file (if {\tt outETreeFile} is of the form {\tt *.etreeb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testFS msglvl msgFile inETreeFile labelflag radius firstEPSfile secondEPSfile
\end{verbatim}
This driver program investigates the storage requirements for a
limited storage forward sparse factorization.
It first reads in a front tree object and for each front $J$, it
determines two quantities:
(1) the amount of in-core storage necessary to factor 
${\widehat J}$ and its boundary, 
and 
(2) the amount of in-core storage necessary to factor 
$J$, $\mbox{par}(J)$, $\mbox{par}^2(J)$, etc.
The program then creates two EPS files, written to {\tt
firstEPSfile} and {\tt secondEPSfile}.
See Figure~\ref{fig-GRD7x7-FStree} for an example.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
If {\tt labelflag = 1}, the node ids are written on the nodes
in the two plots.
\item
Each node will have a circle with radius {\tt radius}.
\item
The {\tt firstEPSfile} 
and {\tt secondEPSfile} 
parameters is the output EPS file for the two plots.
\end{itemize}
\begin{figure}[htbp]
\caption{{\sc GRD7x7}: Working storage for the forward sparse
factorization of the nested dissection ordering.
On the left is the storage required to factor
${\widehat J}$ and its update matrix.
On the right is the storage required to factor
$J$ and all of its ancestors.
Both plots have the same scale.}
\label{fig-GRD7x7-FStree}
\begin{center}
\fbox{
\psfig{file=../../ETree/doc/FS1.eps,height=3.00in,width=3.00in}
}
\fbox{
\psfig{file=../../ETree/doc/FS2.eps,height=3.00in,width=3.00in}
}
\end{center}
\end{figure}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testHeight msglvl msgFile inETreeFile 
\end{verbatim}
This driver program computes the height of the front tree with
respect to factor storage.
This quantity is the minimum amount of working storage 
for a forward sparse factorization.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testIO msglvl msgFile inFile outFile
\end{verbatim}
This driver program reads and writes {\tt ETree} files, useful for
converting formatted files to binary files and vice versa.
One can also read in a {\tt ETree} file and print out just the 
header information (see the {\tt ETree\_writeStats()} method).
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt outFile} parameter is the output file for the {\tt ETree}
object. 
If {\tt outFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to 
a formatted file (if {\tt outFile} is of the form {\tt *.etreef}),
or
a binary file (if {\tt outFile} is of the form {\tt *.etreeb}).
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testMaps msglvl msgFile inETreeFile outIVfile nthread type cutoff
\end{verbatim}
This program is used to construct an owners {\tt IV} that maps a
front to its owning thread or process.
The owners map {\tt IV} object is optionally written to a file.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt outIVFile} parameter is the output file for the 
owners map {\tt IV} object. 
If {\tt outIVFile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outIVFile} 
is of the form {\tt *.ivf}), or a binary file 
(if {\tt outIVFile} is of the form {\tt *.ivb}).
\item
The {\tt nthread} parameter specifies the number of threads or
processes to be used.
\item
The {\tt type} parameter specifies the type of multisector.
\begin{itemize}
\item
{\tt type == 1} 
--- use {\tt ETree\_wrapMap()} to compute a wrap mapping.
\item
{\tt type == 2} 
--- use {\tt ETree\_balancedMap()} to compute a balanced mapping.
\item
{\tt type == 3} 
--- use {\tt ETree\_subtreeSubset()} 
    to compute a subtree-subset mapping.
\item
{\tt type == 4} 
--- use {\tt ETree\_ddMap()} to compute a domain decomposition map.
\end{itemize}
\item
{\tt cutoff} is a cutoff value for the multisector used only for
the domain decomposition map.
The cutoff defines the multisector, $0 \le \mbox{\tt cutoff} \le 1$.
If front {\tt J} has a subtree metric based on forward operations
that is greater than or equalt to {\tt cutoff} times the total
number of operations, then front {\tt J} belongs to the
multisector.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testMS msglvl msgFile inETreeFile outIVfile flag cutoff
\end{verbatim}
This program is used to extract a multisector from a front tree
{\tt ETree} object.
It partitions the vertices into domains and a multisector, where
each domain is a subtree of the elimination tree and the
multisector is the rest of the vertices.
The choice of the subtrees depends on the {\tt flag} and {\tt
cutoff} parameters --- it can be based on depth of a subtree or the
number of vertices, factor entries or factor operations associated
with the subtree.
The component ids {\tt IV} object is optionally written to a file.
Here is some sample output for {\tt BCSSTK30} ordered by nested
dissection, where the multisector is defined by subtree vertex
weight ({\tt flag = 2}) with {\tt cutoff = 0.125}.
\begin{verbatim}
 region  vertices    entries   operations   metric/(avg domain)
     0       1671     597058    255691396  0.797  2.201  3.967
     1       3104     255341     33205237  1.481  0.941  0.515
     2       3222     457255    116441261  1.537  1.685  1.806
     3       1514     194916     41940202  0.722  0.718  0.651
     4       2057     333186    100212056  0.981  1.228  1.555
     5         77       5040       356454  0.037  0.019  0.006
     6       1750     266166     62607526  0.835  0.981  0.971
     7       1887     325977    101994905  0.900  1.202  1.582
     8       3405     492662    125496320  1.624  1.816  1.947
     9       3413     501150    141423868  1.628  1.847  2.194
    10       3242     320220     51679456  1.546  1.180  0.802
    11       2118     238011     44427959  1.010  0.877  0.689
    12       1454     136777     18166107  0.694  0.504  0.282
    13         10        106         1168  0.005  0.000  0.000
 
                    nvtx   %         nzf   %           ops    %
 domains           27253 94.22   3526807 85.52    837952519 76.620
 schur complement   1671  5.78    597058 14.48    255691396 23.380
 total             28924         4123865         1093643915      
\end{verbatim}
Region {\tt 0} is the Schur complement, and there are thirteen
domains, eleven of good size.
A perfectly balanced tree would have eight domains using {\tt cutoff}
equal to 1/8.
It is interesting to see that the Schur complment contains only six
per cent of the vertices but almost one quarter the number of
operations.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\item
The {\tt outIVFile} parameter is the output file for the {\tt IV}
object. 
If {\tt outIVFile} is {\tt none} then the {\tt IV} object is not
written to a file. 
Otherwise, the {\tt IV\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outIVFile} 
is of the form {\tt *.ivf}), or a binary file 
(if {\tt outIVFile} is of the form {\tt *.ivb}).
\item
The {\tt flag} parameter specifies the type of multisector.
\begin{itemize}
\item
{\tt flag == 1} --- the multisector is based on the depth of the
front, i.e., if the front is more than {\tt depth} steps removed
from the root, it forms the root of a domain.
\item
{\tt flag == 2} --- the multisector is based on the number of
vertices in a subtree, i.e., if the subtree rooted at a front
contains more than {\tt cutoff} times the total number of vertices,
it is a domain.
\item
{\tt flag == 3} --- the multisector is based on the number of
factor entries in a subtree, i.e., if the subtree rooted at a front
contains more than {\tt cutoff} times the total number of factor
entries, it is a domain.
\item
{\tt flag == 4} --- the multisector is based on the number of
factor operations in a subtree, i.e., if the subtree rooted at a front
contains more than {\tt cutoff} times the total number of factor
operations, it is a domain.
\end{itemize}
\item
{\tt cutoff} is a cutoff value for the multisector, see above
description when {\tt flag} equals {\tt 1}, {\tt 2} or {\tt 3}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testStats msglvl msgFile inETreeFile labelflag radius firstEPSfile secondEPSfile
\end{verbatim}
This driver program computes one of five metrics associated with a
front tree and writes an EPS file that illustrates the metric overlaid
on the tree structure.
It first reads in a front tree object and 
a graph file.
There are six possible plots:
\begin{center}
\begin{tabular}{c|l}
{\tt metricType} & type of metric \\ \hline
0 & no metric, just a tree plot \\
1 & \# of nodes in a front \\
2 & \# of original matrix entries in a front \\
3 & \# of factor matrix entries in a front \\
4 & \# of forward factor operations in a front \\
5 & \# of backward factor operations in a front 
\end{tabular}
\end{center}
The maximum value of the metric creates a circle with radius
{\tt rmax}, and all other nodes have circles with their area
relative to this largest circle.
See Figure~\ref{fig-GRD7x7x7-metrics} contains four plots,
each used {\tt heightflag = 'D'}, {\tt coordflag = 'P'},
{\tt rmax = 20} and {\tt labelflag = 0}.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\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 outEPSfile} parameter is the name of the EPS file 
to hold the tree.
\item
The {\tt metricType} parameter defines the type of metric to be
illustrated. See above for values.
\item
For information about the {\tt heightflag} and {\tt coordflag}
parameters, see Section~\ref{subsection:Tree:proto:drawing}.
\item
If {\tt labelflag = 1}, the node ids are written on the nodes
in the two plots.
\item
The {\tt fontscale} parameter is the font size when labels are
drawn.
\end{itemize}
\begin{figure}[htbp]
\caption{{\sc GRD7x7x7}: Four tree plots for a $7 \times 7 \times 7$
grid matrix ordered using nested dissection.
The top left tree measure number of original matrix entries in a front.
The top right tree measure number of factor matrix entries in a front.
The bottom left tree measure number of factor operations in a front
for a forward looking factorization, e.g., forward sparse.
The bottom right tree measure number of factor operations in a front
for a backward looking factorization, e.g., general sparse.}
\label{fig-GRD7x7x7-metrics}
\begin{center}
\fbox{
\psfig{file=../../ETree/doc/GRD7x7x7_nzA.eps,height=3.00in,width=3.00in}
}
\fbox{
\psfig{file=../../ETree/doc/GRD7x7x7_nzF.eps,height=3.00in,width=3.00in}
}
\par
\fbox{
\psfig{file=../../ETree/doc/GRD7x7x7_forwops.eps,height=3.00in,width=3.00in}
}
\fbox{
\psfig{file=../../ETree/doc/GRD7x7x7_backops.eps,height=3.00in,width=3.00in}
}
\end{center}
\end{figure}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testStorage msglvl msgFile inETreeFile inGraphFile 
\end{verbatim}
This driver program is used to evaluate the working storage for the
left-looking general sparse and multifrontal algorithms using the
natural post-order traversal of the front tree.
The output is in matlab format to produce a plot.
An example is found below.
\begin{center}
\makebox{
% \psfig{file=workingStorage.eps,width=3.0in,height=2.40in}
\psfig{file=../../ETree/doc/workingStorage.eps,width=3.0in,height=2.40in}
}
\end{center}
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\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.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
testTransform msglvl msgFile inETreeFile inGraphFile 
              outETreeFile maxzeros maxsize seed
\end{verbatim}
This driver program is used to transform a front tree {\tt ETree}
object into a (possibly) merged and (possibly) split front tree.
{\it Merging} the front tree means combining fronts together
that do not introduce more than {\tt maxzeros} zero entries in a
front. (See \cite{ash89-relaxed} and \cite{duf83-multifrontal}
for a description of this supernode amalgamation or relaxation.)
{\it Splitting} a front means breaking a front up into a chain of
smaller fronts; this allows more processors to work on the original
front in a straightforward manner.
The new front tree is optionally written to a file.
Here is some output for the {\tt R3D13824} matrix using
{\tt maxzeros = 1000} and {\tt maxsize = 64}.
\begin{verbatim}
                  CPU  #fronts  #indices  #entries       #ops
 original  :              6001    326858   3459359   1981403337 
 merge one :    0.209     3477    158834   3497139   2000297117 
 merge all :    0.136      748     95306   3690546   2021347776 
 merge any :    0.073      597     85366   3753241   2035158539 
 split     :    0.202      643    115139   3753241   2035158539 
 final     :    3.216      643    115128   3752694   2034396840 
\end{verbatim}
Note how the number of fronts, front indices, factor entries and
factor operations change after each step.
Merging chains (the {\tt merge one} line) halves 
the number of fronts while increasing operations by 1\%.
Merging all children when possible 
(the {\tt merge all} line)
reduces the number of fronts by a factor of 5 
while increasing operations by another 1\%.
Merging any other children (the {\tt merge any} line)
has another additional effect.
Splitting the fronts increases the number of fronts slightly, 
but appears not to change the factor entries or operation counts. 
This is false, as the final step computes the symbolic
factorization for the last front tree and updates the boundary
sizes of the fronts. 
We see that the number of indices, entries and factor operations 
actually decrease slightly due to the split fronts.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output ---
taking {\tt msglvl >= 3} means the {\tt ETree} object is written
to the message file.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt inETreeFile} parameter is the input file for the {\tt ETree}
object. It must be of the form {\tt *.etreef} or {\tt *.etreeb}.
The {\tt ETree} object is read from the file via the
{\tt ETree\_readFromFile()} method.
\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 outETreeFile} parameter is the output file for the {\tt ETree}
object. 
If {\tt outETreeFile} is {\tt none} then the {\tt ETree} object is not
written to a file. 
Otherwise, the {\tt ETree\_writeToFile()} method is called to write
the object to a formatted file (if {\tt outETreeFile} 
is of the form {\tt *.etreef}), or a binary file 
(if {\tt outETreeFile} is of the form {\tt *.etreeb}).
\item
The {\tt maxzeros} parameter is an upper bound on the number of
logically zero entries that will be allowed in a new front.
\item
The {\tt maxsize} parameter is an upper bound on the number of
vertices in a front --- any original front that contains more than
{\tt maxsize} vertices will be broken up into smaller fronts.
\item
{\tt seed} is a seed for a random number generator.
\end{itemize}
%-----------------------------------------------------------------------
\end{enumerate}