Newer
Older
Christian Godenschwager
committed
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
//======================================================================================================================
//
// This file is part of waLBerla. waLBerla is free software: you can
// redistribute it and/or modify it under the terms of the GNU General Public
// License as published by the Free Software Foundation, either version 3 of
// the License, or (at your option) any later version.
//
// waLBerla is distributed in the hope that it will be useful, but WITHOUT
// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
// for more details.
//
// You should have received a copy of the GNU General Public License along
// with waLBerla (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
//
//! \file UniformGrid.cpp
//! \author Florian Schornbaum <florian.schornbaum@fau.de>
//
//======================================================================================================================
#include "blockforest/Initialization.h"
#include "blockforest/SetupBlockForest.h"
#include "blockforest/StructuredBlockForest.h"
#include "blockforest/communication/UniformBufferedScheme.h"
#include "blockforest/communication/UniformDirectScheme.h"
#include "blockforest/loadbalancing/Cartesian.h"
#include "boundary/BoundaryHandling.h"
#include "core/Abort.h"
#include "core/DataTypes.h"
#include "core/cell/CellInterval.h"
#include "core/config/Config.h"
#include "core/debug/Debug.h"
#include "core/debug/TestSubsystem.h"
#include "core/logging/Logging.h"
#include "core/math/IntegerFactorization.h"
#include "core/math/Vector3.h"
#include "core/mpi/Environment.h"
#include "core/mpi/MPIManager.h"
#include "core/mpi/MPIWrapper.h"
#include "core/mpi/Reduce.h"
#include "core/timing/TimingPool.h"
#include "domain_decomposition/SharedSweep.h"
#include "field/AddToStorage.h"
#include "field/FlagField.h"
#include "field/FlagUID.h"
#include "field/communication/PackInfo.h"
#include "field/communication/UniformMPIDatatypeInfo.h"
#include "field/iterators/FieldIterator.h"
#include "field/vtk/FlagFieldCellFilter.h"
#include "field/vtk/VTKWriter.h"
#include "lbm/BlockForestEvaluation.h"
#include "lbm/PerformanceEvaluation.h"
#include "lbm/boundary/NoSlip.h"
#include "lbm/boundary/SimpleUBB.h"
#include "lbm/communication/PdfFieldPackInfo.h"
#include "lbm/communication/PdfFieldMPIDatatypeInfo.h"
#include "lbm/field/AddToStorage.h"
#include "lbm/field/PdfField.h"
#include "lbm/lattice_model/CollisionModel.h"
#include "lbm/lattice_model/D3Q19.h"
#include "lbm/sweeps/CellwiseSweep.h"
#include "lbm/sweeps/SplitPureSweep.h"
#include "lbm/sweeps/SplitSweep.h"
#include "lbm/sweeps/SweepWrappers.h"
#include "lbm/vtk/Density.h"
#include "lbm/vtk/Velocity.h"
#include "postprocessing/sqlite/SQLite.h"
#include "stencil/D3Q19.h"
#include "stencil/D3Q27.h"
#include "timeloop/SweepTimeloop.h"
#include "vtk/BlockCellDataWriter.h"
#include "vtk/Initialization.h"
#include "vtk/VTKOutput.h"
#include <cstdlib>
#include <iostream>
namespace uniform_grid {
///////////
// USING //
///////////
using namespace walberla;
using walberla::uint_t;
using walberla::real_t;
//////////////
// TYPEDEFS //
//////////////
typedef lbm::D3Q19< lbm::collision_model::SRT, false > D3Q19_SRT_INCOMP;
typedef lbm::D3Q19< lbm::collision_model::SRT, true > D3Q19_SRT_COMP;
typedef lbm::D3Q19< lbm::collision_model::TRT, false > D3Q19_TRT_INCOMP;
typedef lbm::D3Q19< lbm::collision_model::TRT, true > D3Q19_TRT_COMP;
typedef lbm::D3Q19< lbm::collision_model::D3Q19MRT, false > D3Q19_MRT_INCOMP;
typedef lbm::D3Q27< lbm::collision_model::D3Q27Cumulant, true > D3Q27_CUMULANT_COMP;
template< typename LatticeModel_T >
struct Types
{
typedef typename LatticeModel_T::Stencil Stencil_T;
typedef typename LatticeModel_T::CommunicationStencil CommunicationStencil_T;
typedef lbm::PdfField< LatticeModel_T > PdfField_T;
};
typedef walberla::uint8_t flag_t;
typedef FlagField< flag_t > FlagField_T;
const uint_t FieldGhostLayers = 1;
///////////
// FLAGS //
///////////
const FlagUID Fluid_Flag( "fluid" );
const FlagUID UBB_Flag( "velocity bounce back" );
const FlagUID NoSlip_Flag( "no slip" );
/////////////////////
// OUTPUT HELPERS //
/////////////////////
template< typename LatticeModel_T, class Enable = void >
struct StencilString;
template< typename LatticeModel_T >
struct StencilString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::Stencil, stencil::D3Q19 >::value >::type >
{
static const char * str() { return "D3Q19"; }
};
template< typename LatticeModel_T >
struct StencilString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::Stencil, stencil::D3Q27 >::value >::type >
{
static const char * str() { return "D3Q27"; }
};
template< typename LatticeModel_T, class Enable = void >
struct CollisionModelString;
template< typename LatticeModel_T >
struct CollisionModelString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::SRT_tag >::value >::type >
{
static const char * str() { return "SRT"; }
};
template< typename LatticeModel_T >
struct CollisionModelString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::TRT_tag >::value >::type >
{
static const char * str() { return "TRT"; }
};
template< typename LatticeModel_T >
struct CollisionModelString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::MRT_tag >::value >::type >
{
static const char * str() { return "MRT"; }
};
template< typename LatticeModel_T >
struct CollisionModelString< LatticeModel_T, typename boost::enable_if_c< boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::Cumulant_tag >::value >::type >
{
static const char * str() { return "Cumulant"; }
};
/////////////
// CONFIG //
/////////////
static inline void getCells( const Config::BlockHandle & configBlock, uint_t & xCells, uint_t & yCells, uint_t & zCells )
{
xCells = configBlock.getParameter< uint_t >( "xCells", 10 );
yCells = configBlock.getParameter< uint_t >( "yCells", 10 );
zCells = configBlock.getParameter< uint_t >( "zCells", 10 );
}
static inline void getCellsAndProcesses( const Config::BlockHandle & configBlock,
const uint_t numberOfProcesses, const uint_t blocksPerProcess,
uint_t & xCells, uint_t & yCells, uint_t & zCells,
uint_t & xProcesses, uint_t & yProcesses, uint_t & zProcesses,
uint_t & xBlocks, uint_t & yBlocks, uint_t & zBlocks )
{
getCells( configBlock, xCells, yCells, zCells );
std::vector< real_t > weighting;
weighting.push_back( configBlock.getParameter< real_t >( "xWeight", real_t(1) ) / real_c(xCells) );
weighting.push_back( configBlock.getParameter< real_t >( "yWeight", real_t(1) ) / real_c(yCells) );
weighting.push_back( configBlock.getParameter< real_t >( "zWeight", real_t(1) ) / real_c(zCells) );
std::vector< uint_t > processes = math::getFactors( numberOfProcesses, 3, weighting );
WALBERLA_CHECK_EQUAL( processes.size(), 3 );
WALBERLA_CHECK_EQUAL( processes[0] * processes[1] * processes[2], numberOfProcesses );
xProcesses = processes[0];
yProcesses = processes[1];
zProcesses = processes[2];
Vector3< uint_t > blocksPerProcess3D = math::getFactors3D( blocksPerProcess );
xBlocks = processes[0] * blocksPerProcess3D[0];
yBlocks = processes[1] * blocksPerProcess3D[1];
zBlocks = processes[2] * blocksPerProcess3D[2];
}
///////////////////////////
// BLOCK STRUCTURE SETUP //
///////////////////////////
void createSetupBlockForest( blockforest::SetupBlockForest & sforest, const Config::BlockHandle & configBlock, const uint_t numberOfProcesses, const uint_t blocksPerProcess )
{
uint_t numberOfXCellsPerBlock, numberOfYCellsPerBlock, numberOfZCellsPerBlock;
uint_t numberOfXProcesses, numberOfYProcesses, numberOfZProcesses;
uint_t numberOfXBlocks, numberOfYBlocks, numberOfZBlocks;
getCellsAndProcesses( configBlock, numberOfProcesses, blocksPerProcess,
numberOfXCellsPerBlock, numberOfYCellsPerBlock, numberOfZCellsPerBlock,
numberOfXProcesses, numberOfYProcesses, numberOfZProcesses,
numberOfXBlocks, numberOfYBlocks, numberOfZBlocks );
sforest.addWorkloadMemorySUIDAssignmentFunction( blockforest::uniformWorkloadAndMemoryAssignment );
sforest.init( AABB( real_t(0), real_t(0), real_t(0), real_c( numberOfXBlocks * numberOfXCellsPerBlock ),
real_c( numberOfYBlocks * numberOfYCellsPerBlock ),
real_c( numberOfZBlocks * numberOfZCellsPerBlock ) ),
numberOfXBlocks, numberOfYBlocks, numberOfZBlocks, false, false, false );
if( MPIManager::instance()->numProcesses() > 1 )
{
shared_ptr< std::vector<uint_t> > processIdMap;
WALBERLA_MPI_SECTION()
{
MPIManager::instance()->createCartesianComm( numberOfXProcesses, numberOfYProcesses, numberOfZProcesses, false, false, false );
processIdMap = make_shared< std::vector<uint_t> >( numberOfXProcesses * numberOfYProcesses * numberOfZProcesses );
for( uint_t z = 0; z != numberOfZProcesses; ++z ) {
for( uint_t y = 0; y != numberOfYProcesses; ++y ) {
for( uint_t x = 0; x != numberOfXProcesses; ++x )
{
(*processIdMap)[z * numberOfXProcesses * numberOfYProcesses + y * numberOfXProcesses + x] =
uint_c( MPIManager::instance()->cartesianRank( x, y, z ) );
}
}
}
}
else
MPIManager::instance()->useWorldComm();
sforest.balanceLoad( blockforest::CartesianDistribution( numberOfXProcesses, numberOfYProcesses, numberOfZProcesses, processIdMap.get() ),
numberOfXProcesses * numberOfYProcesses * numberOfZProcesses );
}
else
{
MPIManager::instance()->useWorldComm();
sforest.balanceLoad( blockforest::CartesianDistribution( numberOfXProcesses, numberOfYProcesses, numberOfZProcesses, NULL ),
numberOfXProcesses * numberOfYProcesses * numberOfZProcesses, real_t(0), 0, true );
}
WALBERLA_LOG_INFO_ON_ROOT( "SetupBlockForest created successfully:\n" << sforest );
}
shared_ptr< blockforest::StructuredBlockForest > createStructuredBlockForest( const Config::BlockHandle & configBlock )
{
uint_t numberOfXCellsPerBlock, numberOfYCellsPerBlock, numberOfZCellsPerBlock;
getCells( configBlock, numberOfXCellsPerBlock, numberOfYCellsPerBlock, numberOfZCellsPerBlock );
const uint_t blocksPerProcess = configBlock.getParameter< uint_t >( "blocksPerProcess", uint_t( 1 ) );
if( configBlock.isDefined( "sbffile" ) )
{
std::string sbffile = configBlock.getParameter< std::string >( "sbffile" );
WALBERLA_LOG_INFO_ON_ROOT( "Creating the block structure: loading from file \'" << sbffile << "\' ..." );
return blockforest::createUniformBlockGrid( sbffile, numberOfXCellsPerBlock, numberOfYCellsPerBlock, numberOfZCellsPerBlock, false );
}
WALBERLA_LOG_INFO_ON_ROOT( "Creating the block structure ..." );
blockforest::SetupBlockForest sforest;
createSetupBlockForest( sforest, configBlock, uint_c( MPIManager::instance()->numProcesses() ), blocksPerProcess );
auto bf = shared_ptr< blockforest::BlockForest >( new blockforest::BlockForest( uint_c( MPIManager::instance()->rank() ), sforest, false ) );
auto sbf = shared_ptr< blockforest::StructuredBlockForest >( new blockforest::StructuredBlockForest( bf, numberOfXCellsPerBlock,
numberOfYCellsPerBlock,
numberOfZCellsPerBlock ) );
sbf->createCellBoundingBoxes();
return sbf;
}
///////////////////////
// BOUNDARY HANDLING //
///////////////////////
template< typename LatticeModel_T >
class MyBoundaryHandling
{
public:
typedef lbm::NoSlip< LatticeModel_T, flag_t > NoSlip_T;
typedef lbm::SimpleUBB< LatticeModel_T, flag_t > UBB_T;
typedef boost::tuples::tuple< NoSlip_T, UBB_T > BoundaryConditions_T;
typedef BoundaryHandling< FlagField_T, typename Types<LatticeModel_T>::Stencil_T, BoundaryConditions_T > BoundaryHandling_T;
MyBoundaryHandling( const BlockDataID & flagField, const BlockDataID & pdfField, const real_t velocity ) :
flagField_( flagField ), pdfField_( pdfField ), velocity_( velocity ) {}
BoundaryHandling_T * operator()( IBlock* const block, const StructuredBlockStorage* const storage ) const;
private:
const BlockDataID flagField_;
const BlockDataID pdfField_;
const real_t velocity_;
}; // class MyBoundaryHandling
template< typename LatticeModel_T >
typename MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T *
MyBoundaryHandling<LatticeModel_T>::operator()( IBlock * const block, const StructuredBlockStorage * const storage ) const
{
typedef typename Types<LatticeModel_T>::PdfField_T PdfField_T;
WALBERLA_ASSERT_NOT_NULLPTR( block );
WALBERLA_ASSERT_NOT_NULLPTR( storage );
FlagField_T * flagField = block->getData< FlagField_T >( flagField_ );
PdfField_T * pdfField = block->getData< PdfField_T > ( pdfField_ );
const auto fluid = flagField->flagExists( Fluid_Flag ) ? flagField->getFlag( Fluid_Flag ) : flagField->registerFlag( Fluid_Flag );
BoundaryHandling_T * handling = new BoundaryHandling_T( "boundary handling", flagField, fluid,
boost::tuples::make_tuple( NoSlip_T( "no slip", NoSlip_Flag, pdfField ),
UBB_T( "velocity bounce back", UBB_Flag, pdfField, velocity_, real_c(0), real_c(0) ) ) );
CellInterval domainBB = storage->getDomainCellBB();
storage->transformGlobalToBlockLocalCellInterval( domainBB, *block );
// no slip WEST
CellInterval west( domainBB.xMin(), domainBB.yMin(), domainBB.zMin(), domainBB.xMin(), domainBB.yMax(), domainBB.zMax() );
handling->forceBoundary( NoSlip_Flag, west );
// no slip EAST
CellInterval east( domainBB.xMax(), domainBB.yMin(), domainBB.zMin(), domainBB.xMax(), domainBB.yMax(), domainBB.zMax() );
handling->forceBoundary( NoSlip_Flag, east );
// no slip SOUTH
CellInterval south( domainBB.xMin(), domainBB.yMin(), domainBB.zMin(), domainBB.xMax(), domainBB.yMin(), domainBB.zMax() );
handling->forceBoundary( NoSlip_Flag, south );
// no slip NORTH
CellInterval north( domainBB.xMin(), domainBB.yMax(), domainBB.zMin(), domainBB.xMax(), domainBB.yMax(), domainBB.zMax() );
handling->forceBoundary( NoSlip_Flag, north );
// no slip BOTTOM
CellInterval bottom( domainBB.xMin(), domainBB.yMin(), domainBB.zMin(), domainBB.xMax(), domainBB.yMax(), domainBB.zMin() );
handling->forceBoundary( NoSlip_Flag, bottom );
// velocity bounce back TOP
CellInterval top( domainBB.xMin(), domainBB.yMin(), domainBB.zMax(), domainBB.xMax(), domainBB.yMax(), domainBB.zMax() );
handling->forceBoundary( UBB_Flag, top );
handling->fillWithDomain( domainBB );
return handling;
}
/////////
// VTK //
/////////
template< typename LatticeModel_T >
class MyVTKOutput {
public:
MyVTKOutput( const ConstBlockDataID & pdfField, const ConstBlockDataID & flagField,
vtk::VTKOutput::BeforeFunction pdfGhostLayerSync ) :
pdfField_( pdfField ), flagField_( flagField ), pdfGhostLayerSync_( pdfGhostLayerSync ) {}
void operator()( std::vector< shared_ptr<vtk::BlockCellDataWriterInterface> > & writers,
std::map< std::string, vtk::VTKOutput::CellFilter > & filters,
std::map< std::string, vtk::VTKOutput::BeforeFunction > & beforeFunctions );
private:
const ConstBlockDataID pdfField_;
const ConstBlockDataID flagField_;
vtk::VTKOutput::BeforeFunction pdfGhostLayerSync_;
}; // class MyVTKOutput
template< typename LatticeModel_T >
void MyVTKOutput<LatticeModel_T>::operator()( std::vector< shared_ptr<vtk::BlockCellDataWriterInterface> > & writers,
std::map< std::string, vtk::VTKOutput::CellFilter > & filters,
std::map< std::string, vtk::VTKOutput::BeforeFunction > & beforeFunctions )
{
// block data writers
writers.push_back( make_shared< lbm::VelocityVTKWriter<LatticeModel_T> >( pdfField_, "VelocityFromPDF" ) );
writers.push_back( make_shared< lbm::DensityVTKWriter<LatticeModel_T> >( pdfField_, "DensityFromPDF" ) );
writers.push_back( make_shared< field::VTKWriter< FlagField_T > >( flagField_, "FlagField" ) );
// cell filters
field::FlagFieldCellFilter<FlagField_T> fluidFilter( flagField_ );
fluidFilter.addFlag( Fluid_Flag );
filters[ "FluidFilter" ] = fluidFilter;
field::FlagFieldCellFilter<FlagField_T> obstacleFilter( flagField_ );
obstacleFilter.addFlag( NoSlip_Flag );
obstacleFilter.addFlag( UBB_Flag );
filters[ "ObstacleFilter" ] = obstacleFilter;
// before functions
beforeFunctions[ "PDFGhostLayerSync" ] = pdfGhostLayerSync_;
}
////////////////////
// THE SIMULATION //
////////////////////
template< typename LatticeModel_T, class Enable = void >
struct AddLB
{
typedef typename Types<LatticeModel_T>::PdfField_T PdfField;
typedef typename Types<LatticeModel_T>::CommunicationStencil_T CommunicationStencil;
static void add( shared_ptr< blockforest::StructuredBlockForest > & blocks, SweepTimeloop & timeloop,
const BlockDataID & pdfFieldId, const BlockDataID & flagFieldId, const BlockDataID & boundaryHandlingId,
const bool split, const bool pure, const bool fullComm, const bool fused, const bool directComm )
{
// setup of the LB communication for synchronizing the pdf field between neighboring blocks
Christian Godenschwager
committed
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
if( directComm )
{
if( fullComm )
{
blockforest::communication::UniformDirectScheme< stencil::D3Q27 > comm( blocks );
auto mpiDatatypeInfo = make_shared<field::communication::UniformMPIDatatypeInfo< PdfField > >( pdfFieldId );
comm.addDataToCommunicate( mpiDatatypeInfo );
commFunction = comm;
}
else
{
blockforest::communication::UniformDirectScheme< CommunicationStencil > comm( blocks );
auto mpiDatatypeInfo = make_shared<lbm::communication::PdfFieldMPIDatatypeInfo< PdfField > >( pdfFieldId );
comm.addDataToCommunicate( mpiDatatypeInfo );
commFunction = comm;
}
}
else
{
if( fullComm )
{
blockforest::communication::UniformBufferedScheme< stencil::D3Q27 > scheme( blocks );
scheme.addPackInfo( make_shared< field::communication::PackInfo< PdfField > >( pdfFieldId ) );
commFunction = scheme;
}
else
{
blockforest::communication::UniformBufferedScheme< CommunicationStencil > scheme( blocks );
scheme.addPackInfo( make_shared< lbm::PdfFieldPackInfo< LatticeModel_T > >( pdfFieldId ) );
commFunction = scheme;
}
}
if( fused )
{
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
if( split )
{
if( pure )
timeloop.add() << Sweep( lbm::SplitPureSweep< LatticeModel_T >( pdfFieldId ), "split pure LB sweep (stream & collide)" );
else
timeloop.add() << Sweep( lbm::SplitSweep< LatticeModel_T, FlagField_T >( pdfFieldId, flagFieldId, Fluid_Flag ), "split LB sweep (stream & collide)" );
}
else
timeloop.add() << Sweep( makeSharedSweep( lbm::makeCellwiseSweep< LatticeModel_T, FlagField_T >( pdfFieldId, flagFieldId, Fluid_Flag ) ), "cell-wise LB sweep (stream & collide)" );
}
else
{
if( split )
{
if( pure )
{
typedef lbm::SplitPureSweep< LatticeModel_T > Sweep_T;
auto sweep = make_shared< Sweep_T >( pdfFieldId );
timeloop.add() << Sweep( lbm::CollideSweep< Sweep_T >( sweep ), "split pure LB sweep (collide)" );
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
timeloop.add() << Sweep( lbm::StreamSweep< Sweep_T >( sweep ), "split pure LB sweep (stream)" );
}
else
{
typedef lbm::SplitSweep< LatticeModel_T, FlagField_T > Sweep_T;
auto sweep = make_shared< Sweep_T >( pdfFieldId, flagFieldId, Fluid_Flag );
timeloop.add() << Sweep( lbm::CollideSweep< Sweep_T >( sweep ), "split LB sweep (collide)" );
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
timeloop.add() << Sweep( lbm::StreamSweep< Sweep_T >( sweep ), "split LB sweep (stream)" );
}
}
else
{
auto sweep = lbm::makeCellwiseSweep< LatticeModel_T, FlagField_T >( pdfFieldId, flagFieldId, Fluid_Flag );
timeloop.add() << Sweep( lbm::makeCollideSweep( sweep ), "cell-wise LB sweep (collide)" );
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
timeloop.add() << Sweep( lbm::makeStreamSweep( sweep ), "cell-wise LB sweep (stream)" );
}
}
}
};
template< typename LatticeModel_T >
struct AddLB< LatticeModel_T, typename boost::enable_if_c< boost::mpl::or_<
boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::MRT_tag >,
boost::is_same< typename LatticeModel_T::CollisionModel::tag,
lbm::collision_model::Cumulant_tag >
>::value >::type >
{
typedef typename Types<LatticeModel_T>::PdfField_T PdfField;
typedef typename Types<LatticeModel_T>::CommunicationStencil_T CommunicationStencil;
static void add( shared_ptr< blockforest::StructuredBlockForest > & blocks, SweepTimeloop & timeloop,
const BlockDataID & pdfFieldId, const BlockDataID & flagFieldId, const BlockDataID & boundaryHandlingId,
const bool /*split*/, const bool /*pure*/, const bool fullComm, const bool fused, const bool directComm )
{
// setup of the LB communication for synchronizing the pdf field between neighboring blocks
Christian Godenschwager
committed
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
if( directComm )
{
if( fullComm )
{
blockforest::communication::UniformDirectScheme< stencil::D3Q27 > comm( blocks );
auto mpiDatatypeInfo = make_shared<field::communication::UniformMPIDatatypeInfo< PdfField > >( pdfFieldId );
comm.addDataToCommunicate( mpiDatatypeInfo );
commFunction = comm;
}
else
{
blockforest::communication::UniformDirectScheme< CommunicationStencil > comm( blocks );
auto mpiDatatypeInfo = make_shared<lbm::communication::PdfFieldMPIDatatypeInfo< PdfField > >( pdfFieldId );
comm.addDataToCommunicate( mpiDatatypeInfo );
commFunction = comm;
}
}
else
{
if( fullComm )
{
blockforest::communication::UniformBufferedScheme< stencil::D3Q27 > scheme( blocks );
scheme.addPackInfo( make_shared< field::communication::PackInfo< PdfField > >( pdfFieldId ) );
commFunction = scheme;
}
else
{
blockforest::communication::UniformBufferedScheme< CommunicationStencil > scheme( blocks );
scheme.addPackInfo( make_shared< lbm::PdfFieldPackInfo< LatticeModel_T > >( pdfFieldId ) );
commFunction = scheme;
}
}
if( fused )
{
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
timeloop.add() << Sweep( makeSharedSweep( lbm::makeCellwiseSweep< LatticeModel_T, FlagField_T >( pdfFieldId, flagFieldId, Fluid_Flag ) ), "cell-wise LB sweep (stream & collide)" );
}
else
{
auto sweep = lbm::makeCellwiseSweep< LatticeModel_T, FlagField_T >( pdfFieldId, flagFieldId, Fluid_Flag );
timeloop.add() << Sweep( lbm::makeCollideSweep( sweep ), "cell-wise LB sweep (collide)" );
timeloop.add() << BeforeFunction( commFunction, "LB communication" )
<< Sweep( MyBoundaryHandling<LatticeModel_T>::BoundaryHandling_T::getBlockSweep( boundaryHandlingId ), "LB boundary sweep" );
timeloop.add() << Sweep( lbm::makeStreamSweep( sweep ), "cell-wise LB sweep (stream)" );
}
}
};
template< typename LatticeModel_T >
void run( const shared_ptr< Config > & config, const LatticeModel_T & latticeModel,
const bool split, const bool pure, const bool fzyx, const bool fullComm, const bool fused, const bool directComm )
{
typedef typename Types<LatticeModel_T>::PdfField_T PdfField;
Config::BlockHandle configBlock = config->getBlock( "UniformGrid" );
// creating the block structure
auto blocks = createStructuredBlockForest( configBlock );
// add pdf field to blocks
BlockDataID pdfFieldId = fzyx ? lbm::addPdfFieldToStorage( blocks, "pdf field (fzyx)", latticeModel,
Vector3< real_t >( real_c(0), real_c(0), real_c(0) ), real_t(1),
FieldGhostLayers, field::fzyx ) :
lbm::addPdfFieldToStorage( blocks, "pdf field (zyxf)", latticeModel,
Vector3< real_t >( real_c(0), real_c(0), real_c(0) ), real_t(1),
FieldGhostLayers, field::zyxf );
// add flag field to blocks
BlockDataID flagFieldId = field::addFlagFieldToStorage< FlagField_T >( blocks, "flag field" );
// add LB boundary handling to blocks
const real_t velocity = configBlock.getParameter< real_t >( "velocity", real_t(0.05) );
BlockDataID boundaryHandlingId = blocks->template addStructuredBlockData< typename MyBoundaryHandling< LatticeModel_T >::BoundaryHandling_T >(
MyBoundaryHandling< LatticeModel_T >( flagFieldId, pdfFieldId, velocity ), "boundary handling" );
// creating the time loop
const uint_t outerTimeSteps = configBlock.getParameter< uint_t >( "outerTimeSteps", uint_c(1 ) );
const uint_t innerTimeSteps = configBlock.getParameter< uint_t >( "innerTimeSteps", uint_c(10) );
SweepTimeloop timeloop( blocks->getBlockStorage(), outerTimeSteps * innerTimeSteps );
// VTK
blockforest::communication::UniformBufferedScheme< stencil::D3Q27 > pdfGhostLayerSync( blocks );
pdfGhostLayerSync.addPackInfo( make_shared< field::communication::PackInfo< PdfField > >( pdfFieldId ) );
MyVTKOutput< LatticeModel_T > myVTKOutput( pdfFieldId, flagFieldId, pdfGhostLayerSync );
std::map< std::string, vtk::SelectableOutputFunction > vtkOutputFunctions;
vtk::initializeVTKOutput( vtkOutputFunctions, myVTKOutput, blocks, config );
for( auto output = vtkOutputFunctions.begin(); output != vtkOutputFunctions.end(); ++output )
timeloop.addFuncBeforeTimeStep( output->second.outputFunction, std::string("VTK: ") + output->first,
output->second.requiredGlobalStates, output->second.incompatibleGlobalStates );
// add LB kernel, boundary handling, and communication to time loop
AddLB< LatticeModel_T >::add( blocks, timeloop, pdfFieldId, flagFieldId, boundaryHandlingId, split, pure, fullComm, fused, directComm );
// logging right before the benchmark starts
lbm::BlockForestEvaluation< FlagField_T > blockForest( blocks, flagFieldId, Fluid_Flag );
blockForest.logInfoOnRoot();
WALBERLA_LOG_INFO_ON_ROOT( "Benchmark parameters:"
"\n- collision model: " << CollisionModelString< LatticeModel_T >::str() <<
"\n- stencil: " << StencilString< LatticeModel_T >::str() <<
"\n- compressible: " << ( LatticeModel_T::compressible ? "yes" : "no" ) <<
"\n- fused (stream & collide) kernel: " << ( fused ? "yes" : "no" ) <<
"\n- split (collision) kernel: " << ( split ? "yes" : "no" ) <<
"\n- pure kernel: " << ( pure ? "yes (collision is also performed within obstacle cells)" : "no" ) <<
"\n- data layout: " << ( fzyx ? "fzyx (structure of arrays [SoA])" : "zyxf (array of structures [AoS])" ) <<
"\n- communication: " << ( fullComm ? "full synchronization" : "direction-aware optimizations" ) <<
"\n- direct communication: " << ( directComm ? "enabled" : "disabled" ) );
// run the benchmark
lbm::PerformanceEvaluation< FlagField_T > performance( blocks, flagFieldId, Fluid_Flag );
for( uint_t outerRun = 0; outerRun < outerTimeSteps; ++outerRun )
{
WcTimingPool timeloopTiming;
WALBERLA_MPI_WORLD_BARRIER();
WcTimer timer;
timer.start();
for( uint_t innerRun = 0; innerRun < innerTimeSteps; ++innerRun )
timeloop.singleStep( timeloopTiming );
timer.end();
double time = timer.max();
mpi::reduceInplace( time, mpi::MAX );
const auto reducedTimeloopTiming = timeloopTiming.getReduced();
WALBERLA_LOG_RESULT_ON_ROOT( "Time loop timing:\n" << *reducedTimeloopTiming );
performance.logResultOnRoot( innerTimeSteps, time );
WALBERLA_ROOT_SECTION()
{
// logging in SQL database
if( configBlock.getParameter< bool >( "logToSqlDB", true ) )
{
const std::string sqlFile = configBlock.getParameter< std::string >( "sqlFile", "performance.sqlite" );
std::map< std::string, int > integerProperties;
std::map< std::string, double > realProperties;
std::map< std::string, std::string > stringProperties;
performance.getResultsForSQLOnRoot( integerProperties, realProperties, stringProperties, innerTimeSteps, time );
blockForest.getResultsForSQLOnRoot( integerProperties, realProperties, stringProperties );
stringProperties[ "collisionModel" ] = CollisionModelString< LatticeModel_T >::str();
stringProperties[ "stencil" ] = StencilString< LatticeModel_T >::str();
stringProperties[ "compressible" ] = ( LatticeModel_T::compressible ? "yes" : "no" );
stringProperties[ "fusedKernel" ] = ( fused ? "yes" : "no" );
stringProperties[ "splitKernel" ] = ( split ? "yes" : "no" );
stringProperties[ "pureKernel" ] = ( pure ? "yes" : "no" );
stringProperties[ "dataLayout" ] = ( fzyx ? "fzyx" : "zyxf" );
stringProperties[ "fullCommunication" ] = ( fullComm ? "yes" : "no" );
stringProperties[ "directComm"] = ( directComm ? "yes" : "no" );
auto runId = postprocessing::storeRunInSqliteDB( sqlFile, integerProperties, stringProperties, realProperties );
postprocessing::storeTimingPoolInSqliteDB( sqlFile, runId, *reducedTimeloopTiming, "Timeloop" );
}
}
}
// logging once again at the end of the simulation, identical to logging at the beginning :-)
blockForest.logInfoOnRoot();
WALBERLA_LOG_INFO_ON_ROOT( "Benchmark parameters:"
"\n- collision model: " << CollisionModelString< LatticeModel_T >::str() <<
"\n- stencil: " << StencilString< LatticeModel_T >::str() <<
"\n- compressible: " << ( LatticeModel_T::compressible ? "yes" : "no" ) <<
"\n- fused (stream & collide) kernel: " << ( fused ? "yes" : "no" ) <<
"\n- split (collision) kernel: " << ( split ? "yes" : "no" ) <<
"\n- pure kernel: " << ( pure ? "yes (collision is also performed within obstacle cells)" : "no" ) <<
"\n- data layout: " << ( fzyx ? "fzyx (structure of arrays [SoA])" : "zyxf (array of structures [AoS])" ) <<
"\n- communication: " << ( fullComm ? "full synchronization" : "direction-aware optimizations" ) <<
"\n- direct communication: " << ( directComm ? "enabled" : "disabled" ) );
}
//////////
// MAIN //
//////////
enum CM { CMSRT, CMTRT, CMMRT, CMCUM };
int main( int argc, char **argv )
{
mpi::Environment env( argc, argv );
if( argc < 2 )
{
WALBERLA_ROOT_SECTION()
{
std::cout << "Usage: " << argv[0] << " path-to-configuration-file [--trt | --mrt] [--comp] [--not-split] [--not-pure] [--zyxf] [--full-comm] [--not-fused] [--direct-comm]\n"
"\n"
"By default, SRT is selected as collision model, a communication with direction-aware optimizations is chosen, and an\n"
"incompressible, split, pure LB kernel is executed on a PDF field with layout 'fzyx' (= structure of arrays [SoA]).\n"
"\n"
"Optional arguments:\n"
" --trt: collision model = TRT\n"
" --mrt: collision model = MRT\n"
" --cumulant collision model = cumulant\n"
" --comp: LB kernel is switched from incompressible to compressible\n"
" --not-split: LB kernel NOT split by PDF direction but executed cell by cell\n"
" --not-pure: LB kernel is only executed in fluid cells, not in obstacle/boundary cells.\n"
" Will be automatically selected for non-split LB kernels.\n"
" --zyxf: data layout switched to 'zyxf' (array of structures [AoS])\n"
" Probably the best layout for non-split kernels.\n"
" --full-comm: A full synchronization of neighboring blocks is performed instead of using a communication\n"
" that uses direction-aware optimizations.\n"
" --not-fused: Selects separate LB kernels for collision and streaming.\n"
" By default, a 'fused' stream & collide kernel is used.\n"
" --direct-comm: Enables bufferless direct communication\n"
"\n"
"Please note: For small/very small blocks (i.e., blocks with only few cells), the best performance may be achieved with\n"
" basic (non-split), incompressible LB kernels combined with an array of structures ('zyxf') data layout!" << std::endl;
}
return EXIT_SUCCESS;
}
logging::Logging::printHeaderOnStream();
//WALBERLA_ROOT_SECTION() { logging::Logging::instance()->setLogLevel( logging::Logging::PROGRESS ); }
#ifdef _OPENMP
if( std::getenv( "OMP_NUM_THREADS" ) == NULL )
WALBERLA_ABORT( "If you are using a version of the benchmark that was compiled with OpenMP you have to "
"specify the environment variable \'OMP_NUM_THREADS\' accordingly!" );
#endif
// open configuration file
shared_ptr< Config > config = make_shared< Config >();
config->readParameterFile( argv[1] );
Config::BlockHandle configBlock = config->getBlock( "UniformGrid" );
if( !configBlock )
WALBERLA_ABORT( "You have to specify a \"UniformGrid\" block in the configuration file!" );
// In case 'sbffile' and 'processes' are specified in the configuration file:
// -> just create the block structure and save it to file for later use
if( configBlock.isDefined( "sbffile" ) && configBlock.isDefined( "processes" ) )
{
std::string sbffile = configBlock.getParameter< std::string >( "sbffile" );
const uint_t numberOfProcesses = configBlock.getParameter< uint_t >( "processes" );
const uint_t blocksPerProcess = configBlock.getParameter< uint_t >( "blocksPerProcess", uint_t( 1 ) );
std::ostringstream infoString;
infoString << "You have selected the option of just creating the block structure (= domain decomposition) and saving the result to file\n"
"by specifying the output file name \'" << sbffile << "\' AND providing a targeted number of processes ("
<< numberOfProcesses << ").\n";
if( MPIManager::instance()->numProcesses() > 1 )
WALBERLA_ABORT( infoString.str() << "In this mode you need to start " << argv[0] << " with just one process!" );
WALBERLA_LOG_INFO_ON_ROOT( infoString.str() << "Creating the block structure ..." );
blockforest::SetupBlockForest sforest;
createSetupBlockForest( sforest, configBlock, numberOfProcesses, blocksPerProcess );
sforest.saveToFile( sbffile.c_str() );
logging::Logging::printFooterOnStream();
return EXIT_SUCCESS;
}
// reading optional parameters from passed arguments
CM collisionModel = CMSRT;
bool compressible = false;
bool split = true;
bool pure = true;
bool fzyx = true;
bool fullComm = false;
bool fused = true;
bool directComm = false;
for( int i = 2; i < argc; ++i )
{
if( std::strcmp( argv[i], "--trt" ) == 0 ) collisionModel = CMTRT;
if( std::strcmp( argv[i], "--mrt" ) == 0 ) collisionModel = CMMRT;
if( std::strcmp( argv[i], "--cumulant" ) == 0 ) collisionModel = CMCUM;
if( std::strcmp( argv[i], "--comp" ) == 0 ) compressible = true;
if( std::strcmp( argv[i], "--not-split" ) == 0 ) split = false;
if( std::strcmp( argv[i], "--not-pure" ) == 0 ) pure = false;
if( std::strcmp( argv[i], "--zyxf" ) == 0 ) fzyx = false;
if( std::strcmp( argv[i], "--full-comm" ) == 0 ) fullComm = true;
if( std::strcmp( argv[i], "--not-fused" ) == 0 ) fused = false;
if( std::strcmp( argv[i], "--direct-comm" ) == 0 ) directComm = true;
}
if( pure && !split )
{
WALBERLA_LOG_WARNING_ON_ROOT( "You called the benchmark with \"--not-split\" but without \"--not-pure\".\n"
"\"Pure\" kernels are only available for \"split\" kernels! Setting \"pure\" to false ..." );
pure = pure && split; // pure only works in combination with split
}
if( collisionModel == CMMRT && ( compressible || split || pure ) )
{
WALBERLA_LOG_WARNING_ON_ROOT( "Option \"--comp\" is not available for MRT! Also, MRT requires options \"--not-split\" and \"--not-pure\"!\n"
"Setting \"compressible\", \"split\", and \"pure\" to false ..." );
compressible = false;
split = false;
pure = false;
}
if( collisionModel == CMCUM && ( compressible || split || pure ) )
{
WALBERLA_LOG_WARNING_ON_ROOT( "Option \"--comp\" has to be set for Cumulant! Also, Cumulant requires options \"--not-split\" and \"--not-pure\"!\n"
"Setting \"compressible\", to true, \"split\", and \"pure\" to false ..." );
compressible = true;
split = false;
pure = false;
}
WALBERLA_NON_MPI_SECTION()
{
if( directComm )
{
WALBERLA_LOG_WARNING_ON_ROOT( "Direct (bufferless) communication is not available when building with MPI disabled.\n"
"Switching to buffered comm..." );
}
directComm = false;
}
const real_t omega = configBlock.getParameter< real_t >( "omega", real_t(1.4) ); // on the coarsest grid!
// executing benchmark
if( collisionModel == CMSRT ) // SRT
{
if( compressible )
{
D3Q19_SRT_COMP latticeModel = D3Q19_SRT_COMP( lbm::collision_model::SRT( omega ) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
else
{
D3Q19_SRT_INCOMP latticeModel = D3Q19_SRT_INCOMP( lbm::collision_model::SRT( omega ) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
}
else if( collisionModel == CMTRT ) // TRT
{
if( compressible )
{
D3Q19_TRT_COMP latticeModel = D3Q19_TRT_COMP( lbm::collision_model::TRT::constructWithMagicNumber( omega ) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
else
{
D3Q19_TRT_INCOMP latticeModel = D3Q19_TRT_INCOMP( lbm::collision_model::TRT::constructWithMagicNumber( omega ) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
}
else if( collisionModel == CMMRT ) // MRT
{
D3Q19_MRT_INCOMP latticeModel = D3Q19_MRT_INCOMP( lbm::collision_model::D3Q19MRT::constructTRTWithMagicNumber( omega ) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
else // Cumulant
{
D3Q27_CUMULANT_COMP latticeModel = D3Q27_CUMULANT_COMP( lbm::collision_model::D3Q27Cumulant(omega) );
run( config, latticeModel, split, pure, fzyx, fullComm, fused, directComm );
}
logging::Logging::printFooterOnStream();
return EXIT_SUCCESS;
}
} // namespace uniform_grid
int main( int argc, char ** argv )