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
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head> <link rel="canonical" href="http://www.mcs.anl.gov/petsc/petsc-current/docs/changes/2024.html" />
<meta http-equiv="content-type" content="text/html;charset=utf-8">
<title>Documentation: Changes: 2.0.24</title>
</head>
<body bgcolor="#ffffff">
<div id="version" align=right><b>petsc-3.4.2 2013-07-02</b></div>
<h1>Documentation: Changes: 2.0.24</h1>
<style type="text/css">
h2 {
color: red;
}
h4 {
font-weight: bold;
text-decoration: underline;
}
</style>
<div id="main">
<div align="center">
<h2><a name="CHANGES">CHANGES</a> in PETSc 2.0.24 (see new <a href="#Features">features</a>)</h2>
</div>
<h4>General:</h4>
<ul>
<li>Changed PETSC_ARCH on Windows platforms from nt to win32, and from nt_gnu to win32_gnu.</li>
<li>
Moved all the compiler/optimization information for each PETSC_ARCH
from the various <code>base.g</code>,
<code>base.O_c++</code> etc. files all into the single file
<code>base_variables</code>. Now configuration for your system
should require only editing the files ${PETSC_ARCH}/base and
${PETSC_ARCH}/<code>base.site</code>
</li>
<li>
The PETSC_ARCH <code>rs6000</code> is now intended for IBM rs6000
workstations running MPICH or LAM, while <code>rs6000_sp</code>
is for IBM SPs.
</li>
<li>
Removed the gvec directory from the distribution. If you are using
this please let us and know and we can provide it or provide
alternative code to do what you need.
</li>
<li>
Routines with the suffix "<code>withcoloring</code>"
have been changed to the suffix "<code>color</code>" to
get around the 31 character limit in names.
</li>
<li>
The source code directory structure has changed slightly: pc and ksp
are now subdirectories of sles, and ao and da are subdirectories of
a new directory called dm (data management).
</li>
<li>
The PETSc libraries are now: -lpetscts -lpetscsnes -lpetscsles
-lpetscdm -lpetscmat -lpetscvec -lpetsc. Each library is now only
dependent on libraries to the right of it in the list above. For
example, if you are only using the linear solvers, you only need to
list "-lpetscsles -lpetscdm -lpetscmat -lpetscvec -lpetsc"
and if you are only using the vectors you only need to list
"-lpetscvec -lpetsc".
</li>
<li>
For users who previously could not use the PETSc makefiles because
the PETSc rules for making .o files conflicted with the user's rules,
we have introduce a new file that may be included by the user,
<code>${PETSC_DIR}/bmake/${PETSC_ARCH}/base_variables</code>,
which defines all the useful PETSc makefile variables, but not the
rules. This should make it easier for people who previously had
trouble using the PETSc makefiles. We highly recommend using the
PETSc makefiles when possible because they are immediately portable
across all systems that support PETSc.
</li>
</ul>
<h4>AO (Application Orderings):</h4>
<h4>TS (Timestepping Solvers):</h4>
<ul>
<li>
Removed <code>TSSetRHSJacobianDefault</code>() - you can now call
<code>TSSetRHSJacobian</code>(...
<code>TSDefaultJacobianColor</code>,....) directly from
Fortran
</li>
<li>Changed <code>TSDefaultJacobianSlow()</code> to <code>TSDefaultJacobian()</code></li>
</ul>
<h4>SNES (Nonlinear Solvers):</h4>
<ul>
<li>
S<code>NESSetConvergenceHistory()</code> takes an additional
argument; see the manual page for details.
</li>
<li>
Changed the interface for matrix-free methods in SNES so that is more
flexible, in particular:
<ul>
<li>
Changed<code> SNESDefaultMatrixFreeMatCreate()</code> to <code>MatSNESFDMFCreate()</code>.
</li>
<li>
Added routines
<ul>
<li>MatSNESFDMFSetHHistory(Mat,Scalar *,int);</li>
<li>MatSNESFDMFResetHHistory(Mat,Scalar *,int);</li>
<li>MatSNESFDMFSetParameters(Mat,double,double);</li>
<li>MatSNESFDMFGetH(Mat,Scalar *);</li>
<li>MatSNESFDMFKSPMonitor(KSP,int,double,void *); and the corresponding option <code>-snes_mf_ksp_monitor</code></li>
</ul>
</li>
<li>
Added support for an alternative way to compute the differencing
parameter in matrix-free methods using the Walker-Pernice
srategy, accessible via the option <code>-snes_mf_type</code> wp
</li>
</ul>
</li>
<li>
Changed the calling sequence of <code>SNESSetLineSearch()</code> so that it accepts an optional user-defined context variable. Likewise, changed the calling sequences of the various line
search routines, e.g., <code>SNESCubicLineSearch(),
SNESQuadraticLineSearch()</code>, etc.
</li>
</ul>
<h4>SLES (Linear Solvers):</h4>
<ul>
<li>See PC and KSP</li>
</ul>
<h4>KSP (Krylov Subspace Methods):</h4>
<ul>
<li>
Changed the calling sequence of
<code>KSPComputeEigenvalues(KSP,int,double*,double*,int
*neig);</code> to return as the final value the number of
eigenvalues computed; see the manual page for more details.
</li>
<li>
Changed GMRES to use the Gram-Schmidt method with one optional step
of iterative refinement as the default orthogonalization
procedure, instead of the modified Gram-Schmidt method. This
change was urged by Rich Lehoucq. It could cause very minor
differences in convergence rates for codes run before and after
this change.
</li>
<li>
The option <code>-ksp_bsmonitor</code> is not longer needed; use
-<code>ksp_truemonitor</code> for the same affect.
</li>
<li>
Added an additional argument to
<code>KSPSetResidualHistory()</code>; see the manual page for
details. This change was suggested by Maria Murillo.
</li>
</ul>
<h4>PC (Preconditioners):</h4>
<ul>
<li>
Removed support for block Gauss-Seidel from the distribution;
<code>-pc_type bgs</code> is no longer supported.
</li>
<li>
The Eisenstat preconditioner (SSOR variant) now uses diagonal scaling
by default, and the options
<code>PCEisenstatUseDiagonalScaling()</code> and
<code>-pc_eisenstat_diagonal_scaling</code> have been
replaced with <code>PCEisenstatNoDiagonalScaling()</code> and
<code>-pc_eisenstat_no_diagonal_scaling</code>
</li>
</ul>
<h4>MAT (Matrices):</h4>
<ul>
<li>Replaced<code> MatGetSubMatrixCall</code>() with <code>MatReuse().</code></li>
<li>
Changed the calling sequence of <code>MatILUFactor()</code> and
<code>MatILUFactorSymbolic()</code> to increase flexibility
of adding new ILU options.
</li>
<li>
Added an additional argument to <code>MatCopy()</code>;
<code>SAME_NONZERO_PATTERN</code> or
<code>DIFFERENT_NONZERO_PATTERN.</code>
</li>
<li>
Added an additional argument to <code>MatDuplicate()</code>,
<code>MAT_COPY_VALUES</code> or
<code>MAT_DO_NOT_COPY_VALUES.</code>
</li>
<li>
Changed the flag <code>MAT_NEW_NONZERO_LOCATION_ERROR</code> to
<code>MAT_NEW_NONZERO_LOCATION_ERR</code>; also changed
<code>MAT_NEW_ALLOCATION_ERROR</code> to
<code>MAT_NEW_ALLOCATION_ERR</code>
</li>
<li>
Changed <code>MatOrder</code>xxx to
<code>MatOrdering</code>xxx; changed
<code>MatOrderingType</code> to char* from enum
</li>
<li>
Changed <code>Mat</code>xxx<code>Reordering</code>xxx to
<code>Mat</code>xxx<code>Ordering</code>xxx changed command
line option <code>-mat_order</code> to
-<code>mat_ordering_type;</code> changed <code>MatColoringType
</code>to char* from enum
</li>
<li>Changed <code>Partitioning</code>xxx to <code>MatPartitioning</code>xxx</li>
</ul>
<h4>DA (Distributed Arrays):</h4>
<ul>
<li>
<code>DAGetInfo()</code> has a new final argument, the stencil
type. See the manual change your current code to work, you only need
add an extra <code>PETSC_NULL</code> as a final argument.
</li>
</ul>
<h4>VEC (Vectors):</h4>
<ul>
<li>
After a call to <code>VecCreate()</code> you MUST either call
<code>VecSetType()</code> or <code>VecSetFromOptions()</code>
on the vector before using it for any purpose. The
<code>VecCreate</code>xxx() routines (e.g.,
<code>VecCreateSeq(), VecCreateMPI()</code>) do not need to have
calls to VecSetType() or VecSetFromOptions() after them.
</li>
<li>
All <code>VecGetArray()</code> calls MUST now be matched with
a <code>VecRestoreArray()</code> call; previous PETSc versions
did not force compliance with this rule.
</li>
<li>
The command line names for changing the vector type in
<code>VecCreate(</code>) are now <code>-vec_type mpi</code>
or <code>-vec_type shared</code> rather than
<code>-vec_mpi</code> or <code>-vec_shared</code>
</li>
<li>Changed the <code>VecType</code> variable type from an enum type to a string.</li>
<li>Changed the calling sequence for <code>VecGetType()</code>; see the manual page for details.</li>
<li>
Changed <code>VecGhostGetLocalRepresentation()</code> and
<code>VecGhostRestoreLocalRepresentation()</code>, to
<code>VecGhostGetLocalForm()</code> and
<code>VecGhostRestoreLocalForm()</code>, respectively; the
previous names had too many characters for some machines.
</li>
</ul>
<h4>IS (Index Sets):</h4>
<h4>Draw (Graphics):</h4>
<ul>
<li>Replaced<code> VIEWER_DRAWX_</code>.... with <code>VIEWER_DRAW_</code>....</li>
<li>
The arguments of <code>ViewerDrawGetDraw</code>() have been
changed; see the manual page for details.
</li>
<li>
<code>DrawTensorContour()</code> has changed; its final argument
is now an array can now use <code>VecView()</code> to create the contour plot.
</li>
</ul>
<h4>Viewers:</h4>
<ul>
<li>
Changed the names associated with the Matlab viewer to Socket viewer;
this reflects our plan to support additional graphics engines with
PETSc.
</li>
<li>Changed <code>ViewerFileOpenASCII()</code> to <code>ViewerASCIIOpen().</code></li>
<li>Changed <code>ViewerDrawOpenX(</code>) to <code>ViewerDrawOpen().</code></li>
<li>Changed <code>ViewerFileOpenBinary()</code> to <code>ViewerBinaryOpen().</code></li>
<li>Changed <code>ViewerType</code> from an enum to a char *.</li>
</ul>
<h4>System:</h4>
<ul>
<li>Changed the calling sequence of <code>PetscGetHomeDirectory(char dir[],int maxlen)</code>.</li>
<li>Changed the calling sequence of <code>PetscFixFilename(const char filein[],char fileout[]).</code></li>
<li>Changed the calling sequence of <code>PetscGetDate().</code></li>
</ul>
<h4>Error Handling:</h4>
<h4>Event Logging:</h4>
<h4>Fortran Interface:</h4>
<div align="center">
<h2>New <a name="Features">features</a> in PETSc 2.0.24 (see <a href="#CHANGES">changes</a> above)</h2>
</div>
<h4>General:</h4>
<ul>
<li>Add the routine <code>OptionsGetAll().</code></li>
<li>Added support for Sun Solaris running on Intel processors; PETSC_ARCH solaris_x86.</li>
<li>
Added option -<code>optionsleft_off</code> so BOPT=g version of
code will not print warnings about unused options.
</li>
<li>The environment variable PETSC_ARCH is now required to use the NT distribution.</li>
<li>Added the routine <code>OptionsGetLogical().</code></li>
<li>We now include PETSC_ARCH of solaris_gnu, rs6000_gnu for those systems when using the gnu compilers.</li>
<li>Thanks to the generosity of the FreeBSD organization, we are now supporting freebsd again.</li>
<li>
Added optional filename to <code>-log_info</code>
<filename> option. PETSc will open a seperate file, called
filename.rank for each processor and print the output there.
</li>
<li>Added <code>PetscTypeCompare()</code>; see the manual page.</li>
</ul>
<h4>AO (Application Orderings):</h4>
<ul>
<li>
<code>AOPetscToApplicationIS()</code> and
<code>AOApplicationToPetscIS()</code>, when passed a stride index
set, now handle them correctly.
</li>
</ul>
<h4>TS (Timestepping Solvers):</h4>
<ul>
<li>Added <code>TSSetTypeFromOptions().</code></li>
</ul>
<h4>SNES (Nonlinear Solvers):</h4>
<ul>
<li>
Added <code>-snes_vecmonitor</code> and
<code>SNESMonitorVec()</code> to allow graphically monitoring the
solution.
</li>
<li>Added <code>SNESSetTypeFromOptions().</code></li>
<li>
Added the routine <code>SNESSetLineSearchCheck()</code>, which
sets a routine to check the validity of new iterates computed by line
search routines; see the manual page for details.
</li>
</ul>
<h4>SLES (Linear Solvers):</h4>
<ul>
<li>See PC and KSP</li>
<li>Added <code>SLESSetTypesFromOptions().</code></li>
</ul>
<h4>KSP (Krylov Subspace Methods):</h4>
<ul>
<li>
Added the function <code>KSPSetAvoidNorms()</code> and
corresponding option <code>-ksp_avoid_norms</code>. For the
methods CG, Richardson, Bi-CG-stab, CR, and CGS, this eliminates the
computation of the residual norm at each iteration (which is an
expensive, collective operation). Then the convergence criterion
depends only on the maximum number of iterations the users sets with
<code>KSPSetTolerances().</code> This is useful when using the
Krylov method as a smoother with a fixed number of iterations (say,
e.g., 3 iterations).
</li>
<li>
Added the KSP type of <code>KSPBICG</code> (contributed by Victor
Eijkhout) and added support to some of the preconditioners to apply
the preconditioner transpose (required by KSPBICG).
</li>
<li>
Added <code>-ksp_vecmonitor</code> and
<code>KSPMonitorVec()</code> to allow graphically monitoring the
solution's convergence.
</li>
<li>Added <code>KSPSetTypeFromOptions().</code></li>
</ul>
<h4>PC (Preconditioners):</h4>
<ul>
<li>
Added <code>PCLUSetReuseReordering()</code> and
<code>PCLUSetReuseFill()</code> so that SLES can reuse the
reordering and fill information for a new factorization of a matrix
with slightly different nonzero structure than the previous
factorization. The corresponding command line options are
-<code>pc_lu_reuse_reordering</code> and
<code>-pc_lu_reuse_fill</code>. These two options only have an
effect when the <code>SLESSetOperators()</code> option of
<code>DIFFERENT_NONZERO_PATTERN</code> is used; otherwise the
ordering and fill information is always reused.
</li>
<li>
Added <code>PCGetVector()</code>, not expected to be used by
application programmers.
</li>
<li>
Added the option <code>-pc_ilu_diagonal_fill</code> and the
corresponding routine <code>PCILUSetAllowDiagonalFill()</code>;
see the manual page for details; helps ILU(k) on matrices with
missing diagonal entries. Suggested by David Hysom.
</li>
<li>Added <code>PCSetTypeFromOptions().</code></li>
</ul>
<h4>MAT (Matrices):</h4>
<ul>
<li>
Added <code>MatSeqAIJSetColumnIndices()</code> and
<code>MatSeqBAIJSetColumnIndices()</code>.
</li>
<li>
If the option <code>MAT_IGNORE_OFF_PROC_ENTRIES</code> is used,
then the routines <code>MatAssemblyBegin()</code> and
<code>MatAssemblyEnd()</code> will not cause any parallel
communication or synchronization.
</li>
<li>
If the option <code>MAT_NEW_NONZERO_LOCATIONS</code> is used, one
less collective call will be made in the
<code>MatAssemblyEnd()</code> routines for MPIAIJ and
MPIBAIJ.
</li>
<li>Fixed bugs in <code>MatReorderForNonzeroDiagonal()</code> and made it more robust.</li>
<li>Fixed bugs in factorization for AIJ matrices with variable size inodes.</li>
<li>
Indices that are negative passed into
<code>MatSetValues...()</code> are now ignored (rather than
generating an error). This allows easily keeping homogeneous
Dirichlet boundary nodes out of the vectors and matrices for
unstructured grid problems; just indicate those node numbers with
negative numbers.
</li>
<li>Added additional optimized code for BAIJ matrices with block sizes 2, 3, 5, 6, and 7.</li>
<li>
Improved the performance of <code>MatSetValues()</code> and
<code>MatAssemblyXXX()</code> for parallel matrices a great
deal.
</li>
<li>
Added the option <code>-matstash_initial_size</code>
<stashsize> and <code>-matstash_block_initial_size</code>
<stashsize> and the routine
<code>MatSetStashInitialSize()</code> to allow further optimization
of off-processor<code>
MatSetValues().</code>
</li>
</ul>
<h4>DA (Distributed Arrays):</h4>
<ul>
<li>Added graphics support to DA vectors in 1d and 2d, with contour plotters for 2d.</li>
</ul>
<h4>VEC (Vectors):</h4>
<ul>
<li>
Extended <code>VecCreate()</code> by allowing registration of new
vector types with <code>VecRegister()</code> see the manual page
for details.
</li>
<li>
Indices that are negative passed into
<code>VecSetValues...()</code> are now ignored (rather than
generating an error). This allows easily keeping homogeneous
Dirichlet boundary nodes out of the vectors and matrices for
unstructured grid problems; just indicate those node numbers with
negative numbers.
</li>
<li>
Added <code>VecDotBegin()</code>, <code>VecDotEnd()</code>,
<code>VecNormBegin()</code>, <code>VecNormEnd()</code>, which
reduce communication overhead in parallel; see the
manual pages for details (suggested by Victor Eijkhout).
</li>
<li>
If the option <code>VEC_IGNORE_OFF_PROC_ENTRIES</code> is used,
the <code>VecAssemblyBegin()</code> and
<code>VecAssemblyEnd()</code> routines will not cause any
parallel communication or synchronization (previously
<code>VecAssemblyBegin()</code> still used a barrier).
</li>
<li>
Added <code>VIEWER_FORMAT_ASCII_INDEX</code> which causes the
index for each vector entry to be printed with the vector
values.
</li>
<li>
Added three new vector scatter options;
<code>-vecscatter_merge</code> and
<code>-vecscatter_ssend</code> and
<code>-vecscatter_sendfirst</code>; these may improve performance
a great deal when using MPICH for MPI on ethernet or fast
ethernet.
</li>
<li>
Added <code>VecStrideNorm(), VecStrideScatter()</code>, and
<code>VecStrideGather().</code>
</li>
<li>
Improved the performance of <code>VecSetValues()</code> and
<code>VecAssemblyXXX()</code> for parallel vectors.
</li>
<li>
Added the option <code>-vecstash_initial_size</code>
<stashsize> and <code>-vecstash_block_initial_size</code>
<stashsize> and the routine
<code>VecSetStashInitialSize()</code> to allow further optimization
of off-processor <code>VecSetValues()</code>.
</li>
</ul>
<h4>IS (Index Sets):</h4>
<ul>
<li>Added <code>ISStrideToGeneral()</code>, which converts a stride type index set to a general type.</li>
</ul>
<h4>Draw (Graphics):</h4>
<h4>Viewers:</h4>
<ul>
<li>
Added <code>ViewerASCIIPushTab()</code>,
<code>ViewerASCIIPopTab()</code> and
<code>ViewerASCIIPrintf()</code>. These allow multi-level
display of solver information (tabing the text further for each inner
solver). Thanks to Victor Eijkhout and Matt Knepley for pushing this.
</li>
</ul>
<h4>System:</h4>
<h4>Error Handling:</h4>
<h4>Event Logging:</h4>
<h4>Fortran Interface:</h4>
<ul>
<li>Added Fortran 90 interface support for HP Convex and Solaris.machine.</li>
</ul>
</div>
</body>
</html>
|
