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
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498 | // *****************************************************************************
/*!
\file src/Inciter/Transporter.cpp
\copyright 2012-2015 J. Bakosi,
2016-2018 Los Alamos National Security, LLC.,
2019-2021 Triad National Security, LLC.
All rights reserved. See the LICENSE file for details.
\brief Transporter drives the time integration of transport equations
\details Transporter drives the time integration of transport equations.
The implementation uses the Charm++ runtime system and is fully asynchronous,
overlapping computation, communication as well as I/O. The algorithm
utilizes the structured dagger (SDAG) Charm++ functionality. The high-level
overview of the algorithm structure and how it interfaces with Charm++ is
discussed in the Charm++ interface file src/Inciter/transporter.ci.
*/
// *****************************************************************************
#include <string>
#include <iostream>
#include <cstddef>
#include <unordered_set>
#include <limits>
#include <cmath>
#include <brigand/algorithms/for_each.hpp>
#include "Macro.hpp"
#include "Transporter.hpp"
#include "Fields.hpp"
#include "PDEStack.hpp"
#include "UniPDF.hpp"
#include "PDFWriter.hpp"
#include "ContainerUtil.hpp"
#include "LoadDistributor.hpp"
#include "MeshReader.hpp"
#include "Inciter/Types.hpp"
#include "Inciter/InputDeck/InputDeck.hpp"
#include "NodeDiagnostics.hpp"
#include "ElemDiagnostics.hpp"
#include "DiagWriter.hpp"
#include "Callback.hpp"
#include "CartesianProduct.hpp"
#include "NoWarning/inciter.decl.h"
#include "NoWarning/partitioner.decl.h"
extern CProxy_Main mainProxy;
namespace inciter {
extern ctr::InputDeck g_inputdeck;
extern ctr::InputDeck g_inputdeck_defaults;
extern std::vector< CGPDE > g_cgpde;
extern std::vector< DGPDE > g_dgpde;
}
using inciter::Transporter;
Transporter::Transporter() :
m_input( input() ),
m_nchare( m_input.size() ),
m_meshid(),
m_ncit( m_nchare.size(), 0 ),
m_nload( 0 ),
m_npart( 0 ),
m_nstat( 0 ),
m_ndisc( 0 ),
m_nchk( 0 ),
m_ncom( 0 ),
m_nt0refit( m_nchare.size(), 0 ),
m_ndtrefit( m_nchare.size(), 0 ),
m_noutrefit( m_nchare.size(), 0 ),
m_noutderefit( m_nchare.size(), 0 ),
m_scheme(),
m_partitioner(),
m_refiner(),
m_meshwriter(),
m_sorter(),
m_nelem( m_nchare.size() ),
m_npoin(),
m_finished( m_nchare.size(), 0 ),
m_meshvol( m_nchare.size() ),
m_minstat( m_nchare.size() ),
m_maxstat( m_nchare.size() ),
m_avgstat( m_nchare.size() ),
m_timer(),
m_progMesh( g_inputdeck.get< tag::cmd, tag::feedback >(),
ProgMeshPrefix, ProgMeshLegend ),
m_progWork( g_inputdeck.get< tag::cmd, tag::feedback >(),
ProgWorkPrefix, ProgWorkLegend )
// *****************************************************************************
// Constructor
// *****************************************************************************
{
// Echo configuration to screen
info( printer() );
const auto nstep = g_inputdeck.get< tag::discr, tag::nstep >();
const auto t0 = g_inputdeck.get< tag::discr, tag::t0 >();
const auto term = g_inputdeck.get< tag::discr, tag::term >();
const auto constdt = g_inputdeck.get< tag::discr, tag::dt >();
// If the desired max number of time steps is larger than zero, and the
// termination time is larger than the initial time, and the constant time
// step size (if that is used) is smaller than the duration of the time to be
// simulated, we have work to do, otherwise, finish right away. If a constant
// dt is not used, that part of the logic is always true as the default
// constdt is zero, see inciter::ctr::InputDeck::InputDeck().
if ( nstep != 0 && term > t0 && constdt < term-t0 ) {
// Enable SDAG waits for collecting mesh statistics
thisProxy.wait4stat();
// Configure and write diagnostics file header
diagHeader();
// Create mesh partitioner AND boundary condition object group
createPartitioner();
} else finish(); // stop if no time stepping requested
}
Transporter::Transporter( CkMigrateMessage* m ) :
CBase_Transporter( m ),
m_progMesh( g_inputdeck.get< tag::cmd, tag::feedback >(),
ProgMeshPrefix, ProgMeshLegend ),
m_progWork( g_inputdeck.get< tag::cmd, tag::feedback >(),
ProgWorkPrefix, ProgWorkLegend )
// *****************************************************************************
// Migrate constructor: returning from a checkpoint
//! \param[in] m Charm++ migrate message
// *****************************************************************************
{
auto print = printer();
print.diag( "Restarted from checkpoint" );
info( print );
inthead( print );
}
std::vector< std::string >
Transporter::input()
// *****************************************************************************
// Generate list of input mesh filenames configured by the user
//! \return List of input mesh filenames configured by the user
//! \details If the input file is given on the command line, a single solver
//! will be instantiated on the single mesh, solving potentially multiple
//! systems of (potentially coupled) equations. If the input file is not given
//! on the command line, the mesh files are expected to be configured in the
//! control/input file, associating a potentially different mesh to each
//! solver. Both configurations allow the solution of coupled systems, but the
//! first one solves all equations on the same mesh, while the latter can
//! couple solutions computed on multiple different meshes.
// *****************************************************************************
{
// Query input mesh filename specified on the command line
const auto& cmdinput = g_inputdeck.get< tag::cmd, tag::io, tag::input >();
// Extract mesh filenames specified in the control file (assigned to solvers)
auto ctrinput = g_inputdeck.mesh();
ErrChk( !cmdinput.empty() || !ctrinput.empty(),
"Either a single input mesh must be given on the command line or multiple "
"meshes must be configured in the control file." );
// Prepend control file path to mesh filenames in given in control file
if (!ctrinput.empty()) {
const auto& ctr = g_inputdeck.get< tag::cmd, tag::io, tag::control >();
auto path = ctr.substr( 0, ctr.find_last_of("/")+1 );
for (auto& f : ctrinput) f = path + f;<--- Consider using std::transform algorithm instead of a raw loop.
}
if (cmdinput.empty()) return ctrinput; else return { cmdinput };
}
void
Transporter::info( const InciterPrint& print )
// *****************************************************************************
// Echo configuration to screen
//! \param[in] print Pretty printer object to use for printing
// *****************************************************************************
{
print.part( "Factory" );
// Print out info data layout
print.list( "Unknowns data layout (CMake: FIELD_DATA_LAYOUT)",
std::list< std::string >{ tk::Fields::layout() } );
// Re-create partial differential equations stack for output
PDEStack stack;
// Print out information on PDE factories
print.eqlist( "Registered PDEs using continuous Galerkin (CG) methods",
stack.cgfactory(), stack.cgntypes() );
print.eqlist( "Registered PDEs using discontinuous Galerkin (DG) methods",
stack.dgfactory(), stack.dgntypes() );
print.endpart();
// Print out information on problem
print.part( "Problem" );
// Print out info on problem title
if ( !g_inputdeck.get< tag::title >().empty() )
print.title( g_inputdeck.get< tag::title >() );
const auto nstep = g_inputdeck.get< tag::discr, tag::nstep >();
const auto t0 = g_inputdeck.get< tag::discr, tag::t0 >();
const auto term = g_inputdeck.get< tag::discr, tag::term >();
const auto constdt = g_inputdeck.get< tag::discr, tag::dt >();
const auto cfl = g_inputdeck.get< tag::discr, tag::cfl >();
const auto scheme = g_inputdeck.get< tag::discr, tag::scheme >();
// Print discretization parameters
print.section( "Discretization parameters" );
print.Item< ctr::Scheme, tag::discr, tag::scheme >();
if (scheme == ctr::SchemeType::DiagCG) {
auto fct = g_inputdeck.get< tag::discr, tag::fct >();
print.item( "Flux-corrected transport (FCT)", fct );
if (fct) {
print.item( "FCT mass diffusion coeff",
g_inputdeck.get< tag::discr, tag::ctau >() );
print.item( "FCT small number",
g_inputdeck.get< tag::discr, tag::fcteps >() );
print.item( "Clipping FCT",
g_inputdeck.get< tag::discr, tag::fctclip >() );
}
} else if (scheme == ctr::SchemeType::DG ||
scheme == ctr::SchemeType::P0P1 || scheme == ctr::SchemeType::DGP1 ||
scheme == ctr::SchemeType::DGP2 || scheme == ctr::SchemeType::PDG)
{
print.Item< ctr::Limiter, tag::discr, tag::limiter >();
}
print.item( "PE-locality mesh reordering",
g_inputdeck.get< tag::discr, tag::pelocal_reorder >() );
print.item( "Operator-access mesh reordering",
g_inputdeck.get< tag::discr, tag::operator_reorder >() );
auto steady = g_inputdeck.get< tag::discr, tag::steady_state >();
print.item( "Local time stepping", steady );
if (steady) {
print.item( "L2-norm residual convergence criterion",
g_inputdeck.get< tag::discr, tag::residual >() );
print.item( "Convergence criterion component index",
g_inputdeck.get< tag::discr, tag::rescomp >() );
}
print.item( "Number of time steps", nstep );
print.item( "Start time", t0 );
print.item( "Terminate time", term );
if (std::abs(constdt - g_inputdeck_defaults.get< tag::discr, tag::dt >()) >
std::numeric_limits< tk::real >::epsilon())
print.item( "Constant time step size", constdt );
else if (std::abs(cfl - g_inputdeck_defaults.get< tag::discr, tag::cfl >()) >
std::numeric_limits< tk::real >::epsilon())
{
print.item( "CFL coefficient", cfl );
}
// Print out info on settings of selected partial differential equations
print.pdes( "Partial differential equations integrated", stack.info() );
// Print out solver coupling info
print.couple( g_inputdeck.get< tag::couple, tag::transfer >(),
g_inputdeck.depvar() );
// Print out adaptive polynomial refinement configuration
if (scheme == ctr::SchemeType::PDG) {
print.section( "Polynomial refinement (p-ref)" );
print.item( "p-refinement",
g_inputdeck.get< tag::pref, tag::pref >() );
print.Item< ctr::PrefIndicator, tag::pref, tag::indicator >();
print.item( "Max degrees of freedom",
g_inputdeck.get< tag::pref, tag::ndofmax >() );
print.item( "Tolerance",
g_inputdeck.get< tag::pref, tag::tolref >() );
}
// Print out adaptive mesh refinement configuration
const auto amr = g_inputdeck.get< tag::amr, tag::amr >();
if (amr) {
print.section( "Mesh refinement (h-ref)" );
print.refvar( g_inputdeck.get< tag::amr, tag::refvar >(),
g_inputdeck.get< tag::amr, tag::id >() );
print.Item< ctr::AMRError, tag::amr, tag::error >();
auto t0ref = g_inputdeck.get< tag::amr, tag::t0ref >();
print.item( "Refinement at t<0 (t0ref)", t0ref );
if (t0ref) {
const auto& initref = g_inputdeck.get< tag::amr, tag::init >();
print.item( "Initial refinement steps", initref.size() );
print.edgeref( g_inputdeck.get< tag::amr, tag::edge >() );
auto eps =
std::numeric_limits< kw::amr_xminus::info::expect::type >::epsilon();
auto xminus = g_inputdeck.get< tag::amr, tag::xminus >();
auto xminus_default = g_inputdeck_defaults.get< tag::amr, tag::xminus >();
if (std::abs( xminus - xminus_default ) > eps)
print.item( "Initial refinement x-", xminus );
auto xplus = g_inputdeck.get< tag::amr, tag::xplus >();
auto xplus_default = g_inputdeck_defaults.get< tag::amr, tag::xplus >();
if (std::abs( xplus - xplus_default ) > eps)
print.item( "Initial refinement x+", xplus );
auto yminus = g_inputdeck.get< tag::amr, tag::yminus >();
auto yminus_default = g_inputdeck_defaults.get< tag::amr, tag::yminus >();
if (std::abs( yminus - yminus_default ) > eps)
print.item( "Initial refinement y-", yminus );
auto yplus = g_inputdeck.get< tag::amr, tag::yplus >();
auto yplus_default = g_inputdeck_defaults.get< tag::amr, tag::yplus >();
if (std::abs( yplus - yplus_default ) > eps)
print.item( "Initial refinement y+", yplus );
auto zminus = g_inputdeck.get< tag::amr, tag::zminus >();
auto zminus_default = g_inputdeck_defaults.get< tag::amr, tag::zminus >();
if (std::abs( zminus - zminus_default ) > eps)
print.item( "Initial refinement z-", zminus );
auto zplus = g_inputdeck.get< tag::amr, tag::zplus >();
auto zplus_default = g_inputdeck_defaults.get< tag::amr, tag::zplus >();
if (std::abs( zplus - zplus_default ) > eps)
print.item( "Initial refinement z+", zplus );
}
auto dtref = g_inputdeck.get< tag::amr, tag::dtref >();
print.item( "Refinement at t>0 (dtref)", dtref );
if (dtref) {
auto dtfreq = g_inputdeck.get< tag::amr, tag::dtfreq >();
print.item( "Mesh refinement frequency, t>0", dtfreq );
print.item( "Uniform-only mesh refinement, t>0",
g_inputdeck.get< tag::amr, tag::dtref_uniform >() );
}
print.item( "Refinement tolerance",
g_inputdeck.get< tag::amr, tag::tolref >() );
print.item( "De-refinement tolerance",
g_inputdeck.get< tag::amr, tag::tolderef >() );
}
// Print out ALE configuration
const auto ale = g_inputdeck.get< tag::ale, tag::ale >();
if (ale) {
print.section( "Arbitrary Lagrangian-Eulerian (ALE) mesh motion" );
auto dvcfl = g_inputdeck.get< tag::ale, tag::dvcfl >();
print.item( "Volume-change CFL coefficient", dvcfl );
print.Item< ctr::MeshVelocity, tag::ale, tag::meshvelocity >();
print.Item< ctr::MeshVelocitySmoother, tag::ale, tag::smoother >();
print.item( "Mesh motion dimensions", tk::parameters(
g_inputdeck.get< tag::ale, tag::mesh_motion >() ) );
const auto& meshforce = g_inputdeck.get< tag::ale, tag::meshforce >();
print.item( "Mesh velocity force coefficients", tk::parameters(meshforce) );
print.item( "Vorticity multiplier",
g_inputdeck.get< tag::ale, tag::vortmult >() );
print.item( "Mesh velocity linear solver tolerance",
g_inputdeck.get< tag::ale, tag::tolerance >() );
print.item( "Mesh velocity linear solver maxit",
g_inputdeck.get< tag::ale, tag::maxit >() );
const auto& dir = g_inputdeck.get< tag::ale, tag::bcdir >();
if (not dir.empty())
print.item( "Mesh velocity Dirichlet BC sideset(s)",
tk::parameters( dir ) );
const auto& sym = g_inputdeck.get< tag::ale, tag::bcsym >();
if (not sym.empty())
print.item( "Mesh velocity symmetry BC sideset(s)",
tk::parameters( sym ) );
std::size_t i = 1;
for (const auto& m : g_inputdeck.get< tag::ale, tag::move >()) {
tk::ctr::UserTable opt;
print.item( opt.group() + ' ' + std::to_string(i) + " interpreted as",
opt.name( m.get< tag::fntype >() ) );
const auto& s = m.get< tag::sideset >();
if (not s.empty())
print.item( "Moving sideset(s) with table " + std::to_string(i),
tk::parameters(s));
++i;
}
}
// Print I/O filenames
print.section( "Input/Output filenames and directories" );
print.item( "Input mesh(es)", tk::parameters( m_input ) );
const auto& of = g_inputdeck.get< tag::cmd, tag::io, tag::output >();
print.item( "Volume field output file(s)",
of + ".e-s.<meshid>.<numchares>.<chareid>" );
print.item( "Surface field output file(s)",
of + "-surf.<surfid>.e-s.<meshid>.<numchares>.<chareid>" );
print.item( "History output file(s)", of + ".hist.{pointid}" );
print.item( "Diagnostics file",
g_inputdeck.get< tag::cmd, tag::io, tag::diag >() );
print.item( "Checkpoint/restart directory",
g_inputdeck.get< tag::cmd, tag::io, tag::restart >() + '/' );
// Print output intervals
print.section( "Output intervals (in units of iteration count)" );
print.item( "TTY", g_inputdeck.get< tag::output, tag::iter, tag::tty>() );
print.item( "Field and surface",
g_inputdeck.get< tag::output, tag::iter, tag::field >() );
print.item( "History",
g_inputdeck.get< tag::output, tag::iter, tag::history >() );
print.item( "Diagnostics",
g_inputdeck.get< tag::output, tag::iter, tag::diag >() );
print.item( "Checkpoint/restart",
g_inputdeck.get< tag::cmd, tag::rsfreq >() );
auto tf = g_inputdeck.get< tag::output, tag::time, tag::field >();
auto th = g_inputdeck.get< tag::output, tag::time, tag::history >();
if (tf>0.0 || th>0.0) {
print.section( "Output intervals (in units of physics time)" );
if (tf > 0.0) print.item( "Field and surface", tf );
if (th > 0.0) print.item( "History", th );
}
const auto& rf = g_inputdeck.get< tag::output, tag::range, tag::field >();
const auto& rh = g_inputdeck.get< tag::output, tag::range, tag::history >();
if (not rf.empty() or not rh.empty()) {
print.section( "Output time ranges (in units of physics time)" );
for (const auto& r : rf)
print.item("Field output { mintime, maxtime, dt }", tk::parameters(r));
for (const auto& r : rh)
print.item("History output { mintime, maxtime, dt }", tk::parameters(r));
}
// Print output variables: fields and surfaces
const auto nodeoutvars = g_inputdeck.outvars( tk::Centering::NODE );
const auto elemoutvars = g_inputdeck.outvars( tk::Centering::ELEM );
const auto aliases = g_inputdeck.outvar_aliases();
const auto outsets = g_inputdeck.outsets();
if (!nodeoutvars.empty() || !elemoutvars.empty() || !outsets.empty())
print.section( "Output fields" );
if (!nodeoutvars.empty())
print.item( "Node field(s)", tk::parameters(nodeoutvars) );
if (!elemoutvars.empty())
print.item( "Elem field(s)", tk::parameters(elemoutvars) );
if (!aliases.empty())
print.item( "Alias(es)", tk::parameters(aliases) );
if (!outsets.empty())
print.item( "Surface side set(s)", tk::parameters(outsets) );
// Print output variables: history
const auto& pt = g_inputdeck.get< tag::history, tag::point >();
const auto& id = g_inputdeck.get< tag::history, tag::id >();
if (!pt.empty()) {
print.section( "Output time history" );
for (std::size_t p=0; p<pt.size(); ++p) {
std::stringstream ss;
auto prec = g_inputdeck.get< tag::prec, tag::history >();
ss << std::setprecision( static_cast<int>(prec) );
ss << of << ".hist." << id[p];
print.longitem( "At point " + id[p] + ' ' + tk::parameters(pt[p]),
ss.str() );
}
}
print.endsubsection();
}
bool
Transporter::matchBCs( std::map< int, std::vector< std::size_t > >& bnd )
// *****************************************************************************
// Verify that side sets specified in the control file exist in mesh file
//! \details This function does two things: (1) it verifies that the side
//! sets used in the input file (either to which boundary conditions (BC)
//! are assigned or listed as field output by the user in the
//! input file) all exist among the side sets read from the input mesh
//! file and errors out if at least one does not, and (2) it matches the
//! side set ids at which the user has configured BCs (or listed as an output
//! surface) to side set ids read from the mesh file and removes those face
//! and node lists associated to side sets that the user did not set BCs or
//! listed as field output on (as they will not need processing further since
//! they will not be used).
//! \param[in,out] bnd Node or face lists mapped to side set ids
//! \return True if sidesets have been used and found in mesh
// *****************************************************************************
{
using PDETypes = ctr::parameters::Keys;
// Query side set ids at which BCs assigned for all BC types for all PDEs
using PDEsBCs =
tk::cartesian_product< PDETypes, ctr::bc::Keys >;
std::unordered_set< int > usedsets;
brigand::for_each< PDEsBCs >( UserBC( g_inputdeck, usedsets ) );
// Query side sets of time dependent BCs (since tag::bctimedep is not a part
// of tag::bc)
brigand::for_each< PDETypes >( UserTimedepBC(g_inputdeck, usedsets) );
// Query side sets of boundaries prescribed as moving with ALE
for (const auto& move : g_inputdeck.get< tag::ale, tag::move >())
for (auto i : move.get< tag::sideset >())
usedsets.insert( std::stoi(i) );
// Add sidesets requested for field output
const auto& ss = g_inputdeck.get< tag::cmd, tag::io, tag::surface >();
for (const auto& s : ss) {
std::stringstream conv( s );
int num;
conv >> num;
usedsets.insert( num );
}
// Find user-configured side set ids among side sets read from mesh file
std::unordered_set< int > sidesets_used;
for (auto i : usedsets) { // for all side sets used in control file
if (bnd.find(i) != end(bnd)) // used set found among side sets in file
sidesets_used.insert( i ); // store side set id configured as BC
else {
Throw( "Boundary conditions specified on side set " +
std::to_string(i) + " which does not exist in mesh file" );
}
}
// Remove sidesets not used (will not process those further)
tk::erase_if( bnd, [&]( auto& item ) {
return sidesets_used.find( item.first ) == end(sidesets_used);
});
return !bnd.empty();
}
void
Transporter::createPartitioner()
// *****************************************************************************
// Create mesh partitioner AND boundary conditions group
// *****************************************************************************
{
auto scheme = g_inputdeck.get< tag::discr, tag::scheme >();
auto centering = ctr::Scheme().centering( scheme );<--- Variable 'centering' is assigned a value that is never used.
auto print = printer();
// Create partitioner callbacks (order important)
tk::PartitionerCallback cbp {{
CkCallback( CkReductionTarget(Transporter,load), thisProxy )
, CkCallback( CkReductionTarget(Transporter,partitioned), thisProxy )
, CkCallback( CkReductionTarget(Transporter,distributed), thisProxy )
, CkCallback( CkReductionTarget(Transporter,refinserted), thisProxy )
, CkCallback( CkReductionTarget(Transporter,refined), thisProxy )
}};
// Create refiner callbacks (order important)
tk::RefinerCallback cbr {{
CkCallback( CkReductionTarget(Transporter,queriedRef), thisProxy )
, CkCallback( CkReductionTarget(Transporter,respondedRef), thisProxy )
, CkCallback( CkReductionTarget(Transporter,compatibility), thisProxy )
, CkCallback( CkReductionTarget(Transporter,bndint), thisProxy )
, CkCallback( CkReductionTarget(Transporter,matched), thisProxy )
, CkCallback( CkReductionTarget(Transporter,refined), thisProxy )
}};
// Create sorter callbacks (order important)
tk::SorterCallback cbs {{
CkCallback( CkReductionTarget(Transporter,queried), thisProxy )
, CkCallback( CkReductionTarget(Transporter,responded), thisProxy )
, CkCallback( CkReductionTarget(Transporter,discinserted), thisProxy )
, CkCallback( CkReductionTarget(Transporter,workinserted), thisProxy )
}};
// Start timer measuring preparation of mesh(es) for partitioning
m_timer[ TimerTag::MESH_READ ];
// Start preparing mesh(es)
print.diag( "Reading mesh(es)" );
// Create (discretization) Scheme chare worker arrays for all meshes
for ([[maybe_unused]] const auto& filename : m_input)
m_scheme.emplace_back( g_inputdeck.get< tag::discr, tag::scheme >(),
g_inputdeck.get< tag::ale, tag::ale >(),
need_linearsolver() );
ErrChk( !m_input.empty(), "No input mesh" );
// Read boundary (side set) data from a list of input mesh files
std::size_t meshid = 0;
for (const auto& filename : m_input) {
// Create mesh reader for reading side sets from file
tk::MeshReader mr( filename );
// Read out total number of mesh points from mesh file
m_npoin.push_back( mr.npoin() );
std::map< int, std::vector< std::size_t > > bface;
std::map< int, std::vector< std::size_t > > faces;
std::map< int, std::vector< std::size_t > > bnode;
// Read boundary-face connectivity on side sets
mr.readSidesetFaces( bface, faces );
bool bcs_set = false;
if (centering == tk::Centering::ELEM) {
// Verify boundarty condition (BC) side sets used exist in mesh file
bcs_set = matchBCs( bface );
} else if (centering == tk::Centering::NODE) {
// Read node lists on side sets
bnode = mr.readSidesetNodes();
// Verify boundarty condition (BC) side sets used exist in mesh file
bcs_set = matchBCs( bnode );
bcs_set = bcs_set || matchBCs( bface );
}
// Warn on no BCs
if (!bcs_set) print << "\n>>> WARNING: No boundary conditions set\n\n";
auto opt = m_scheme[meshid].arrayoptions();
// Create empty mesh refiner chare array (bound to workers)
m_refiner.push_back( CProxy_Refiner::ckNew(opt) );
// Create empty mesh sorter Charm++ chare array (bound to workers)
m_sorter.push_back( CProxy_Sorter::ckNew(opt) );
// Create MeshWriter chare group for mesh
m_meshwriter.push_back(
tk::CProxy_MeshWriter::ckNew(
g_inputdeck.get< tag::selected, tag::filetype >(),
centering,
g_inputdeck.get< tag::cmd, tag::benchmark >(),
m_input.size() ) );
// Create mesh partitioner Charm++ chare nodegroup for all meshes
m_partitioner.push_back(
CProxy_Partitioner::ckNew( meshid, filename, cbp, cbr, cbs,
thisProxy, m_refiner.back(), m_sorter.back(), m_meshwriter.back(),
m_scheme, bface, faces, bnode ) );
++meshid;
}
}
void
Transporter::load( std::size_t meshid, std::size_t nelem )
// *****************************************************************************
// Reduction target: the mesh has been read from file on all PEs
//! \param[in] meshid Mesh id (summed accross all compute nodes)
//! \param[in] nelem Number of mesh elements per mesh (summed across all
//! compute nodes)
// *****************************************************************************
{
meshid /= static_cast< std::size_t >( CkNumNodes() );
Assert( meshid < m_nelem.size(), "MeshId indexing out" );
m_nelem[meshid] = nelem;
// Compute load distribution given total work (nelem) and user-specified
// virtualization
uint64_t chunksize, remainder;
m_nchare[meshid] = static_cast<int>(
tk::linearLoadDistributor(
g_inputdeck.get< tag::cmd, tag::virtualization >(),
m_nelem[meshid], CkNumPes(), chunksize, remainder ) );
// Store sum of meshids (across all chares, key) for each meshid (value).
// This is used to look up the mesh id after collectives that sum their data.
m_meshid[ static_cast<std::size_t>(m_nchare[meshid])*meshid ] = meshid;
Assert( meshid < m_nelem.size(), "MeshId indexing out" );
// Partition first mesh
if (meshid == 0) m_partitioner[0].partition( m_nchare[0] );
if (++m_nload == m_nelem.size()) { // all meshes have been loaded
m_nload = 0;
auto print = printer();
// Start timer measuring preparation of the mesh for partitioning
const auto& timer = tk::cref_find( m_timer, TimerTag::MESH_READ );
print.diag( "Mesh read time: " + std::to_string( timer.dsec() ) + " sec" );
// Print out mesh partitioning configuration
print.section( "Mesh partitioning" );
print.Item< tk::ctr::PartitioningAlgorithm,
tag::selected, tag::partitioner >();
print.item( "Virtualization [0.0...1.0]",
g_inputdeck.get< tag::cmd, tag::virtualization >() );
// Print out initial mesh statistics
meshstat( "Initial load distribution" );
// Tell meshwriter the total number of chares
m_meshwriter[meshid].nchare( m_nchare[meshid] );
// Query number of initial mesh refinement steps
int nref = 0;
if (g_inputdeck.get< tag::amr, tag::t0ref >())
nref = static_cast<int>(g_inputdeck.get< tag::amr, tag::init >().size());
m_progMesh.start( print, "Preparing mesh", {{ CkNumPes(), CkNumPes(), nref,
m_nchare[0], m_nchare[0], m_nchare[0], m_nchare[0] }} );
}
}
void
Transporter::partitioned( std::size_t meshid )
// *****************************************************************************
// Reduction target: a mesh has been partitioned
//! \param[in] meshid Mesh id
// *****************************************************************************
{
if (++m_npart == m_nelem.size()) { // all meshes have been partitioned
m_npart = 0;
} else // partition next mesh
m_partitioner[meshid+1].partition( m_nchare[meshid+1] );
}
void
Transporter::distributed( std::size_t meshid )
// *****************************************************************************
// Reduction target: all compute nodes have distributed their mesh after
// partitioning
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_partitioner[meshid].refine();
}
void
Transporter::refinserted( std::size_t meshid, std::size_t error )
// *****************************************************************************
// Reduction target: all compute nodes have created the mesh refiners
//! \param[in] meshid Mesh id (aggregated across all compute nodes with operator
//! max)
//! \param[in] error Error code (aggregated across all compute nodes with
//! operator max)
// *****************************************************************************
{
if (error) {
printer() << "\n>>> ERROR: A worker chare was not assigned any mesh "
"elements after distributing mesh " + std::to_string(meshid) +
". This can happen in SMP-mode with a large +ppn "
"parameter (number of worker threads per logical node) and is "
"most likely the fault of the mesh partitioning algorithm not "
"tolerating the case when it is asked to divide the "
"computational domain into a number of partitions different "
"than the number of ranks it is called on, i.e., in case of "
"overdecomposition and/or calling the partitioner in SMP mode "
"with +ppn larger than 1. Solution 1: Try a different "
"partitioning algorithm (e.g., rcb instead of mj). Solution 2: "
"Decrease +ppn.";
finish( meshid );
} else {
m_refiner[meshid].doneInserting();
}
}
void
Transporter::queriedRef( std::size_t meshid )
// *****************************************************************************
// Reduction target: all Refiner chares have queried their boundary edges
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_refiner[meshid].response();
}
void
Transporter::respondedRef( std::size_t meshid )
// *****************************************************************************
// Reduction target: all Refiner chares have setup their boundary edges
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_refiner[meshid].refine();
}
void
Transporter::compatibility( std::size_t meshid, std::size_t modified )
// *****************************************************************************
// Reduction target: all Refiner chares have received a round of edges,
// and have run their compatibility algorithm
//! \param[in] meshid Mesh id (aggregated across all chares using operator max)
//! \param[in] modified Modified flag, aggregated across all chares using
//! operator max), if nonzero, mesh is modified
//! \details This is called iteratively, until convergence by Refiner. At this
//! point all Refiner chares have received a round of edge data (tags whether
//! an edge needs to be refined, etc.), and applied the compatibility
//! algorithm independent of other Refiner chares. We keep going until the
//! mesh is no longer modified by the compatibility algorithm, based on a new
//! round of edge data communication started in Refiner::comExtra().
// *****************************************************************************
{
if (modified)
m_refiner[meshid].comExtra();
else
m_refiner[meshid].correctref();
}
void
Transporter::matched( std::size_t summeshid,
std::size_t nextra,
std::size_t nref,
std::size_t nderef,
std::size_t sumrefmode )
// *****************************************************************************
// Reduction target: all Refiner chares have matched/corrected the tagging
// of chare-boundary edges, all chares are ready to perform refinement.
//! \param[in] summeshid Mesh id (summed across all chares)
//! \param[in] nextra Sum (across all chares) of the number of edges on each
//! chare that need correction along chare boundaries
//! \param[in] nref Sum of number of refined tetrahedra across all chares.
//! \param[in] nderef Sum of number of derefined tetrahedra across all chares.
//! \param[in] sumrefmode Sum of contributions from all chares, encoding
//! refinement mode of operation.
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, summeshid );
// If at least a single edge on a chare still needs correction, do correction,
// otherwise, this mesh refinement step is complete
if (nextra > 0) {
++m_ncit[meshid];
m_refiner[meshid].comExtra();
} else {
auto print = printer();
// decode refmode
auto refmode = static_cast< Refiner::RefMode >(
sumrefmode / static_cast<std::size_t>(m_nchare[meshid]) );
if (refmode == Refiner::RefMode::T0REF) {
if (!g_inputdeck.get< tag::cmd, tag::feedback >()) {
ctr::AMRInitial opt;
print.diag( { "meshid", "t0ref", "type", "nref", "nderef", "ncorr" },
{ std::to_string(meshid),
std::to_string(m_nt0refit[meshid]),
"initial",
std::to_string(nref),
std::to_string(nderef),
std::to_string(m_ncit[meshid]) } );
++m_nt0refit[meshid];
}
m_progMesh.inc< REFINE >( print );
} else if (refmode == Refiner::RefMode::DTREF) {
auto dtref_uni = g_inputdeck.get< tag::amr, tag::dtref_uniform >();
print.diag( { "meshid", "dtref", "type", "nref", "nderef", "ncorr" },
{ std::to_string(meshid),
std::to_string(++m_ndtrefit[meshid]),
(dtref_uni?"uniform":"error"),
std::to_string(nref),
std::to_string(nderef),
std::to_string(m_ncit[meshid]) },
false );
} else if (refmode == Refiner::RefMode::OUTREF) {
print.diag( { "meshid", "outref", "nref", "nderef", "ncorr" },
{ std::to_string(meshid),
std::to_string(++m_noutrefit[meshid]),
std::to_string(nref),
std::to_string(nderef),
std::to_string(m_ncit[meshid]) }, false );
} else if (refmode == Refiner::RefMode::OUTDEREF) {
print.diag( { "meshid", "outderef", "nref", "nderef", "ncorr" },
{ std::to_string(meshid),
std::to_string(++m_noutderefit[meshid]),
std::to_string(nref),
std::to_string(nderef),
std::to_string(m_ncit[meshid]) },
false );
} else Throw( "RefMode not implemented" );
m_ncit[meshid] = 0;
m_refiner[meshid].perform();
}
}
void
Transporter::bndint( tk::real sx, tk::real sy, tk::real sz, tk::real cb,
tk::real summeshid )
// *****************************************************************************
// Compute surface integral across the whole problem and perform leak-test
//! \param[in] sx X component of vector summed
//! \param[in] sy Y component of vector summed
//! \param[in] sz Z component of vector summed
//! \param[in] cb Invoke callback if positive
//! \param[in] summeshid Mesh id (summed accross all chares)
//! \details This function aggregates partial surface integrals across the
//! boundary faces of the whole problem. After this global sum a
//! non-zero vector result indicates a leak, e.g., a hole in the boundary,
//! which indicates an error in the boundary face data structures used to
//! compute the partial surface integrals.
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, static_cast<std::size_t>(summeshid) );
std::stringstream err;
if (cb < 0.0) {
err << "Mesh boundary leaky after mesh refinement step; this is due to a "
"problem with updating the side sets used to specify boundary conditions "
"on faces: ";
} else if (cb > 0.0) {
err << "Mesh boundary leaky during initialization; this is due to "
"incorrect or incompletely specified boundary conditions for a given input "
"mesh: ";
}
auto eps = 1.0e-10;
if (std::abs(sx) > eps || std::abs(sy) > eps || std::abs(sz) > eps) {
err << "Integral result must be a zero vector: " << std::setprecision(12) <<
std::abs(sx) << ", " << std::abs(sy) << ", " << std::abs(sz) <<
", eps = " << eps;
Throw( err.str() );
}
if (cb > 0.0) m_scheme[meshid].bcast< Scheme::resizeComm >();
}
void
Transporter::refined( std::size_t summeshid,
std::size_t nelem,
std::size_t npoin )
// *****************************************************************************
// Reduction target: all chares have refined their mesh
//! \param[in] summeshid Mesh id (summed accross all Refiner chares)
//! \param[in] nelem Total number of elements in mesh summed across the
//! distributed mesh
//! \param[in] npoin Total number of mesh points summed across the distributed
//! mesh. Note that in parallel this is larger than the number of points in
//! the mesh, because the boundary nodes are multi-counted. But we only need
//! an equal or larger than npoin for Sorter::setup, so this is okay.
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, summeshid );
// Store new number of elements for initially refined mesh
m_nelem[meshid] = nelem;
m_sorter[meshid].doneInserting();
m_sorter[meshid].setup( npoin );
}
void
Transporter::queried( std::size_t meshid )
// *****************************************************************************
// Reduction target: all Sorter chares have queried their boundary edges
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_sorter[meshid].response();
}
void
Transporter::responded( std::size_t meshid )
// *****************************************************************************
// Reduction target: all Sorter chares have responded with their boundary edges
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_sorter[meshid].start();
}
void
Transporter::resized( std::size_t meshid )
// *****************************************************************************
// Reduction target: all worker chares have resized their own mesh data after
// AMR or ALE
//! \param[in] meshid Mesh id
//! \note Only used for nodal schemes
// *****************************************************************************
{
m_scheme[meshid].disc().vol();
m_scheme[meshid].bcast< Scheme::lhs >();
}
void
Transporter::startEsup( std::size_t meshid )
// *****************************************************************************
// Reduction target: all worker chares have generated their own esup
//! \param[in] meshid Mesh id
//! \note Only used for cell-centered schemes
// *****************************************************************************
{
m_scheme[meshid].bcast< Scheme::nodeNeighSetup >();
}
void
Transporter::discinserted( std::size_t meshid )
// *****************************************************************************
// Reduction target: all Discretization chares have been inserted
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_scheme[meshid].disc().doneInserting();
}
void
Transporter::meshstat( const std::string& header ) const
// *****************************************************************************
// Print out mesh statistics
//! \param[in] header Section header
// *****************************************************************************
{
auto print = printer();
print.section( header );
if (m_nelem.size() > 1) {
print.item( "Number of tetrahedra (per mesh)",tk::parameters(m_nelem) );
print.item( "Number of points (per mesh)", tk::parameters(m_npoin) );
print.item( "Number of work units (per mesh)", tk::parameters(m_nchare) );
}
print.item( "Total number of tetrahedra",
std::accumulate( begin(m_nelem), end(m_nelem), 0UL ) );
print.item( "Total number of points",
std::accumulate( begin(m_npoin), end(m_npoin), 0UL ) );
print.item( "Total number of work units",
std::accumulate( begin(m_nchare), end(m_nchare), 0 ) );
print.endsubsection();
}
bool
Transporter::need_linearsolver() const
// *****************************************************************************
// Decide if we need a linear solver for ALE
//! \return True if ALE will neeed a linear solver
// *****************************************************************************
{
auto smoother = g_inputdeck.get< tag::ale, tag::smoother >();
if ( g_inputdeck.get< tag::ale, tag::ale >() and
(smoother == ctr::MeshVelocitySmootherType::LAPLACE ||
smoother == ctr::MeshVelocitySmootherType::HELMHOLTZ) )
{
return true;
} else {
return false;
}
}
void
Transporter::disccreated( std::size_t summeshid, std::size_t npoin )
// *****************************************************************************
// Reduction target: all Discretization constructors have been called
//! \param[in] summeshid Mesh id (summed accross all chares)
//! \param[in] npoin Total number of mesh points (summed across all chares)
//! Note that as opposed to npoin in refined(), this npoin is not
//! multi-counted, and thus should be correct in parallel.
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, summeshid );
// Update number of mesh points for mesh, since it may have been refined
if (g_inputdeck.get< tag::amr, tag::t0ref >()) m_npoin[meshid] = npoin;
if (++m_ndisc == m_nelem.size()) { // all Disc arrays have been created
m_ndisc = 0;
auto print = printer();
m_progMesh.end( print );
if (g_inputdeck.get< tag::amr, tag::t0ref >())
meshstat( "Initially (t<0) refined mesh graph statistics" );
}
m_refiner[meshid].sendProxy();
if (g_inputdeck.get< tag::discr, tag::scheme >() == ctr::SchemeType::DiagCG)
m_scheme[meshid].fct().doneInserting();
if (g_inputdeck.get< tag::ale, tag::ale >())
m_scheme[meshid].ale().doneInserting();
if (need_linearsolver())
m_scheme[meshid].conjugategradients().doneInserting();
m_scheme[meshid].disc().vol();
}
void
Transporter::workinserted( std::size_t meshid )
// *****************************************************************************
// Reduction target: all worker (derived discretization) chares have been
// inserted
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_scheme[meshid].bcast< Scheme::doneInserting >();
}
void
Transporter::diagHeader()
// *****************************************************************************
// Configure and write diagnostics file header
// *****************************************************************************
{
// Output header for diagnostics output file
tk::DiagWriter dw( g_inputdeck.get< tag::cmd, tag::io, tag::diag >(),
g_inputdeck.get< tag::flformat, tag::diag >(),
g_inputdeck.get< tag::prec, tag::diag >() );
// Collect variables names for integral/diagnostics output
std::vector< std::string > var;
const auto scheme = g_inputdeck.get< tag::discr, tag::scheme >();
if (scheme == ctr::SchemeType::DiagCG || scheme == ctr::SchemeType::ALECG)
for (const auto& eq : g_cgpde) varnames( eq, var );
else if (scheme == ctr::SchemeType::DG ||
scheme == ctr::SchemeType::P0P1 || scheme == ctr::SchemeType::DGP1 ||
scheme == ctr::SchemeType::DGP2 || scheme == ctr::SchemeType::PDG)
for (const auto& eq : g_dgpde) varnames( eq, var );
else Throw( "Diagnostics header not handled for discretization scheme" );
const tk::ctr::Error opt;
auto nv = var.size();
std::vector< std::string > d;
// Add 'L2(var)' for all variables as those are always computed
const auto& l2name = opt.name( tk::ctr::ErrorType::L2 );
for (std::size_t i=0; i<nv; ++i) d.push_back( l2name + '(' + var[i] + ')' );
// Query user-requested diagnostics and augment diagnostics file header by
// 'err(var)', where 'err' is the error type configured, and var is the
// variable computed, for all variables and all error types configured.
const auto& err = g_inputdeck.get< tag::diag, tag::error >();
for (const auto& e : err) {
const auto& errname = opt.name( e );
for (std::size_t i=0; i<nv; ++i)
d.push_back( errname + '(' + var[i] + "-IC)" );
}
// Augment diagnostics variables by L2-norm of the residual and total energy
if (scheme == ctr::SchemeType::DiagCG || scheme == ctr::SchemeType::ALECG) {
for (std::size_t i=0; i<nv; ++i) d.push_back( "L2(d" + var[i] + ')' );
}
d.push_back( "mE" );
// Write diagnostics header
dw.header( d );
}
void
Transporter::comfinal( std::size_t initial, std::size_t summeshid )
// *****************************************************************************
// Reduction target indicating that communication maps have been setup
//! \param[in] initial Sum of contributions from all chares. If larger than
//! zero, we are during time stepping and if zero we are during setup.
//! \param[in] summeshid Mesh id (summed accross the distributed mesh)
// *****************************************************************************
// [Discretization-specific communication maps]
{
auto meshid = tk::cref_find( m_meshid, static_cast<std::size_t>(summeshid) );
if (initial > 0) {
m_scheme[meshid].bcast< Scheme::setup >();
// Turn on automatic load balancing
if (++m_ncom == m_nelem.size()) { // all worker arrays have finished
m_ncom = 0;
auto print = printer();
m_progWork.end( print );
tk::CProxy_LBSwitch::ckNew();
print.diag( "Load balancing on (if enabled in Charm++)" );
}
} else {
m_scheme[meshid].bcast< Scheme::lhs >();
}
}
// [Discretization-specific communication maps]
void
Transporter::totalvol( tk::real v, tk::real initial, tk::real summeshid )
// *****************************************************************************
// Reduction target summing total mesh volume across all workers
//! \param[in] v Mesh volume summed across the distributed mesh
//! \param[in] initial Sum of contributions from all chares. If larger than
//! zero, we are during setup, if zero, during time stepping.
//! \param[in] summeshid Mesh id (summed accross the distributed mesh)
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, static_cast<std::size_t>(summeshid) );
m_meshvol[meshid] = v;
if (initial > 0.0) // during initialization
m_scheme[meshid].disc().stat( v );
else // during ALE or AMR
m_scheme[meshid].bcast< Scheme::resized >();
}
void
Transporter::minstat( tk::real d0, tk::real d1, tk::real d2, tk::real rmeshid )
// *****************************************************************************
// Reduction target yielding minimum mesh statistcs across all workers
//! \param[in] d0 Minimum mesh statistics collected over all chares
//! \param[in] d1 Minimum mesh statistics collected over all chares
//! \param[in] d2 Minimum mesh statistics collected over all chares
//! \param[in] rmeshid Mesh id as a real
// *****************************************************************************
{
auto meshid = static_cast<std::size_t>(rmeshid);
m_minstat[meshid][0] = d0; // minimum edge length
m_minstat[meshid][1] = d1; // minimum cell volume cubic root
m_minstat[meshid][2] = d2; // minimum number of cells on chare
minstat_complete(meshid);
}
void
Transporter::maxstat( tk::real d0, tk::real d1, tk::real d2, tk::real rmeshid )
// *****************************************************************************
// Reduction target yielding the maximum mesh statistics across all workers
//! \param[in] d0 Maximum mesh statistics collected over all chares
//! \param[in] d1 Maximum mesh statistics collected over all chares
//! \param[in] d2 Maximum mesh statistics collected over all chares
//! \param[in] rmeshid Mesh id as a real
// *****************************************************************************
{
auto meshid = static_cast<std::size_t>(rmeshid);
m_maxstat[meshid][0] = d0; // maximum edge length
m_maxstat[meshid][1] = d1; // maximum cell volume cubic root
m_maxstat[meshid][2] = d2; // maximum number of cells on chare
maxstat_complete(meshid);
}
void
Transporter::sumstat( tk::real d0, tk::real d1, tk::real d2, tk::real d3,
tk::real d4, tk::real d5, tk::real summeshid )
// *****************************************************************************
// Reduction target yielding the sum mesh statistics across all workers
//! \param[in] d0 Sum mesh statistics collected over all chares
//! \param[in] d1 Sum mesh statistics collected over all chares
//! \param[in] d2 Sum mesh statistics collected over all chares
//! \param[in] d3 Sum mesh statistics collected over all chares
//! \param[in] d4 Sum mesh statistics collected over all chares
//! \param[in] d5 Sum mesh statistics collected over all chares
//! \param[in] summeshid Mesh id (summed accross the distributed mesh)
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, static_cast<std::size_t>(summeshid) );
m_avgstat[meshid][0] = d1 / d0; // average edge length
m_avgstat[meshid][1] = d3 / d2; // average cell volume cubic root
m_avgstat[meshid][2] = d5 / d4; // average number of cells per chare
sumstat_complete(meshid);
}
void
Transporter::pdfstat( CkReductionMsg* msg )
// *****************************************************************************
// Reduction target yielding PDF of mesh statistics across all workers
//! \param[in] msg Serialized PDF
// *****************************************************************************
{
std::size_t meshid;
std::vector< tk::UniPDF > pdf;
// Deserialize final PDF
PUP::fromMem creator( msg->getData() );
creator | meshid;<--- Uninitialized variable: meshid
creator | pdf;
delete msg;
auto id = std::to_string(meshid);<--- Uninitialized variable: meshid
// Create new PDF file (overwrite if exists)
tk::PDFWriter pdfe( "mesh_edge_pdf." + id + ".txt" );
// Output edgelength PDF
// cppcheck-suppress containerOutOfBounds
pdfe.writeTxt( pdf[0],
tk::ctr::PDFInfo{ {"PDF"}, {}, {"edgelength"}, 0, 0.0 } );
// Create new PDF file (overwrite if exists)
tk::PDFWriter pdfv( "mesh_vol_pdf." + id + ".txt" );
// Output cell volume cubic root PDF
pdfv.writeTxt( pdf[1],<--- Access out of bounds
tk::ctr::PDFInfo{ {"PDF"}, {}, {"V^{1/3}"}, 0, 0.0 } );
// Create new PDF file (overwrite if exists)
tk::PDFWriter pdfn( "mesh_ntet_pdf." + id + ".txt" );
// Output number of cells PDF
pdfn.writeTxt( pdf[2],<--- Access out of bounds
tk::ctr::PDFInfo{ {"PDF"}, {}, {"ntets"}, 0, 0.0 } );
pdfstat_complete(meshid);
}
void
Transporter::stat()
// *****************************************************************************
// Echo diagnostics on mesh statistics
// *****************************************************************************
{
auto print = printer();
if (++m_nstat == m_nelem.size()) { // stats from all meshes have arrived
m_nstat = 0;
for (std::size_t i=0; i<m_nelem.size(); ++i) {
print.diag(
"Mesh " + std::to_string(i) +
" distribution statistics: min/max/avg(edgelength) = " +
std::to_string( m_minstat[i][0] ) + " / " +
std::to_string( m_maxstat[i][0] ) + " / " +
std::to_string( m_avgstat[i][0] ) + ", " +
"min/max/avg(V^{1/3}) = " +
std::to_string( m_minstat[i][1] ) + " / " +
std::to_string( m_maxstat[i][1] ) + " / " +
std::to_string( m_avgstat[i][1] ) + ", " +
"min/max/avg(ntets) = " +
std::to_string( static_cast<std::size_t>(m_minstat[i][2]) ) + " / " +
std::to_string( static_cast<std::size_t>(m_maxstat[i][2]) ) + " / " +
std::to_string( static_cast<std::size_t>(m_avgstat[i][2]) ) );
}
// Print out time integration header to screen
inthead( print );
m_progWork.start( print, "Preparing workers",
{{ m_nchare[0], m_nchare[0], m_nchare[0], m_nchare[0], m_nchare[0] }} );
// Create "derived-class" workers
for (std::size_t i=0; i<m_nelem.size(); ++i) m_sorter[i].createWorkers();
}
}
void
Transporter::boxvol( tk::real v, tk::real summeshid )
// *****************************************************************************
// Reduction target computing total volume of IC box
//! \param[in] v Total volume within user-specified box IC
//! \param[in] summeshid Mesh id as a real (summed accross the distributed mesh)
// *****************************************************************************
{
auto meshid = tk::cref_find( m_meshid, static_cast<std::size_t>(summeshid) );
if (v > 0.0) printer().diag( "Box IC volume: " + std::to_string(v) );
m_scheme[meshid].bcast< Scheme::box >( v );
}
void
Transporter::inthead( const InciterPrint& print )
// *****************************************************************************
// Print out time integration header to screen
//! \param[in] print Pretty printer object to use for printing
// *****************************************************************************
{
auto refined = g_inputdeck.get< tag::cmd, tag::io, tag::refined >();
const auto scheme = g_inputdeck.get< tag::discr, tag::scheme >();
if (refined && scheme == ctr::SchemeType::DG) {
printer() << "\n>>> WARNING: Ignoring refined field output for DG(P0)\n\n";
refined = false;
}
print.inthead( "Time integration", "Navier-Stokes solver",
"Legend: it - iteration count\n"
" t - physics time\n"
" dt - physics time step size\n"
" ETE - estimated wall-clock time elapsed (h:m:s)\n"
" ETA - estimated wall-clock time for accomplishment (h:m:s)\n"
" EGT - estimated grind wall-clock time (ms/timestep)\n"
" flg - status flags, legend:\n"
" f - " + std::string(refined ? "refined " : "")
+ "field (volume and surface)\n"
" d - diagnostics\n"
" t - physics time history\n"
" h - h-refinement\n"
" l - load balancing\n"
" r - checkpoint\n"
" a - ALE mesh velocity linear solver did not converge\n",
"\n it t dt ETE ETA EGT flg\n"
" -------------------------------------------------------------------------\n" );
}
void
Transporter::diagnostics( CkReductionMsg* msg )
// *****************************************************************************
// Reduction target optionally collecting diagnostics, e.g., residuals
//! \param[in] msg Serialized diagnostics vector aggregated across all PEs
//! \note Only used for nodal schemes
// *****************************************************************************
{
std::size_t meshid;
std::vector< std::vector< tk::real > > d;
// Deserialize diagnostics vector
PUP::fromMem creator( msg->getData() );
creator | meshid;<--- Uninitialized variable: meshid
creator | d;
delete msg;
auto id = std::to_string(meshid);<--- Uninitialized variable: meshid<--- Variable 'id' is assigned a value that is never used.
auto ncomp = g_inputdeck.get< tag::component >().nprop();
Assert( ncomp > 0, "Number of scalar components must be positive");
Assert( d.size() == NUMDIAG, "Diagnostics vector size mismatch" );
for (std::size_t i=0; i<d.size(); ++i)<--- Unsigned less than zero
Assert( d[i].size() == ncomp, "Size mismatch at final stage of "
"diagnostics aggregation for mesh " + id );
// Allocate storage for those diagnostics that are always computed
std::vector< tk::real > diag( ncomp, 0.0 );
// Finish computing diagnostics
for (std::size_t i=0; i<d[L2SOL].size(); ++i)
diag[i] = sqrt( d[L2SOL][i] / m_meshvol[meshid] );<--- Uninitialized variable: meshid
// Query user-requested error types to output
const auto& error = g_inputdeck.get< tag::diag, tag::error >();
decltype(ncomp) n = 0;<--- Variable 'n' is assigned a value that is never used.
for (const auto& e : error) {
n += ncomp;<--- Variable 'n' is assigned a value that is never used.
if (e == tk::ctr::ErrorType::L2) {
// Finish computing the L2 norm of the numerical - analytical solution
for (std::size_t i=0; i<d[L2ERR].size(); ++i)
diag.push_back( sqrt( d[L2ERR][i] / m_meshvol[meshid] ) );<--- Uninitialized variable: meshid
} else if (e == tk::ctr::ErrorType::LINF) {
// Finish computing the Linf norm of the numerical - analytical solution
for (std::size_t i=0; i<d[LINFERR].size(); ++i)
diag.push_back( d[LINFERR][i] );
}
}
// Finish computing the L2 norm of the residual and append
const auto scheme = g_inputdeck.get< tag::discr, tag::scheme >();
std::vector< tk::real > l2res( d[L2RES].size(), 0.0 );
if (scheme == ctr::SchemeType::DiagCG || scheme == ctr::SchemeType::ALECG) {
for (std::size_t i=0; i<d[L2RES].size(); ++i) {
l2res[i] = std::sqrt( d[L2RES][i] / m_meshvol[meshid] );<--- Uninitialized variable: meshid
diag.push_back( l2res[i] );
}
}
// Append total energy
diag.push_back( d[TOTALSOL][0] );
// Append diagnostics file at selected times
auto filename = g_inputdeck.get< tag::cmd, tag::io, tag::diag >();
if (m_nelem.size() > 1) filename += '.' + id;
tk::DiagWriter dw( filename,
g_inputdeck.get< tag::flformat, tag::diag >(),
g_inputdeck.get< tag::prec, tag::diag >(),
std::ios_base::app );
dw.diag( static_cast<uint64_t>(d[ITER][0]), d[TIME][0], d[DT][0], diag );
// Continue time step
m_scheme[meshid].bcast< Scheme::refine >( l2res );<--- Uninitialized variable: meshid
}
void
Transporter::resume()
// *****************************************************************************
// Resume execution from checkpoint/restart files
//! \details This is invoked by Charm++ after the checkpoint is done, as well as
//! when the restart (returning from a checkpoint) is complete
// *****************************************************************************
{
if (std::any_of(begin(m_finished), end(m_finished), [](auto f){return !f;})) {
// If just restarted from a checkpoint, Main( CkMigrateMessage* msg ) has
// increased nrestart in g_inputdeck, but only on PE 0, so broadcast.
auto nrestart = g_inputdeck.get< tag::cmd, tag::io, tag::nrestart >();
for (std::size_t i=0; i<m_nelem.size(); ++i)
m_scheme[i].bcast< Scheme::evalLB >( nrestart );
} else
mainProxy.finalize();
}
void
Transporter::checkpoint( std::size_t finished, std::size_t meshid )
// *****************************************************************************
// Save checkpoint/restart files
//! \param[in] finished Nonzero if finished with time stepping
//! \param[in] meshid Mesh id
// *****************************************************************************
{
m_finished[meshid] = finished;
if (++m_nchk == m_nelem.size()) { // all worker arrays have checkpointed
m_nchk = 0;
#ifndef HAS_EXAM2M
const auto benchmark = g_inputdeck.get< tag::cmd, tag::benchmark >();
if (!benchmark) {
const auto& restart = g_inputdeck.get< tag::cmd, tag::io, tag::restart >();
CkCallback res( CkIndex_Transporter::resume(), thisProxy );
CkStartCheckpoint( restart.c_str(), res );
} else {
resume();
}
#else
resume();
#endif
}
}
void
Transporter::finish( std::size_t meshid )
// *****************************************************************************
// Normal finish of time stepping
//! \param[in] meshid Mesh id
// *****************************************************************************
{
checkpoint( /* finished = */ 1, meshid );
}
#include "NoWarning/transporter.def.h"
|