Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
waLBerla
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Model registry
Analyze
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
Michael Zikeli
waLBerla
Commits
e92b3e96
Commit
e92b3e96
authored
4 years ago
by
Helen Schottenhamml
Browse files
Options
Downloads
Patches
Plain Diff
First test for non uniform buffered communication with refinement PackInfo.
parent
881f2cd4
Branches
Branches containing commit
Tags
Tags containing commit
No related merge requests found
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
tests/blockforest/CMakeLists.txt
+5
-0
5 additions, 0 deletions
tests/blockforest/CMakeLists.txt
tests/blockforest/communication/NonUniformBufferedSchemeTest.cpp
+414
-0
414 additions, 0 deletions
...lockforest/communication/NonUniformBufferedSchemeTest.cpp
with
419 additions
and
0 deletions
tests/blockforest/CMakeLists.txt
+
5
−
0
View file @
e92b3e96
...
...
@@ -51,3 +51,8 @@ waLBerla_compile_test( FILES communication/DirectionBasedReduceCommTest.cpp DEPE
waLBerla_execute_test
(
NAME DirectionBasedReduceCommTest1 COMMAND $<TARGET_FILE:DirectionBasedReduceCommTest>
)
waLBerla_execute_test
(
NAME DirectionBasedReduceCommTest3 COMMAND $<TARGET_FILE:DirectionBasedReduceCommTest> PROCESSES 3
)
waLBerla_execute_test
(
NAME DirectionBasedReduceCommTest8 COMMAND $<TARGET_FILE:DirectionBasedReduceCommTest> PROCESSES 8
)
waLBerla_compile_test
(
FILES communication/NonUniformBufferedSchemeTest.cpp DEPENDS blockforest field
)
waLBerla_execute_test
(
NAME NonUniformBufferedSchemeTest1 COMMAND $<TARGET_FILE:NonUniformBufferedSchemeTest>
)
waLBerla_execute_test
(
NAME NonUniformBufferedSchemeTest3 COMMAND $<TARGET_FILE:NonUniformBufferedSchemeTest> PROCESSES 3
)
waLBerla_execute_test
(
NAME NonUniformBufferedSchemeTest8 COMMAND $<TARGET_FILE:NonUniformBufferedSchemeTest> PROCESSES 8
)
\ No newline at end of file
This diff is collapsed.
Click to expand it.
tests/blockforest/communication/NonUniformBufferedSchemeTest.cpp
0 → 100644
+
414
−
0
View file @
e92b3e96
//======================================================================================================================
//
// 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 NonUniformBufferedSchemeTest.cpp
//! \ingroup comm
//! \author Helen Schottenhamml <helen.schottenhamml@fau.de>
//! \brief Checks communication for non-uniform buffered communication with field::refinement::PackInfo
//
//======================================================================================================================
#include
<blockforest/communication/NonUniformBufferedScheme.h>
#include
<blockforest/SetupBlockForest.h>
#include
<blockforest/StructuredBlockForest.h>
#include
<blockforest/loadbalancing/StaticCurve.h>
#include
<boundary/BoundaryHandling.h>
#include
<core/DataTypes.h>
#include
<core/debug/TestSubsystem.h>
#include
<core/logging/Logging.h>
#include
<core/math/Limits.h>
#include
<core/math/Sample.h>
#include
<core/mpi/Environment.h>
#include
<field/AddToStorage.h>
#include
<field/iterators/FieldIterator.h>
#include
<field/refinement/all.h>
#include
<stencil/all.h>
#include
<cstdlib>
#include
<functional>
namespace
nonuniform_buffered_scheme_test
{
///////////
// USING //
///////////
using
namespace
walberla
;
using
walberla
::
real_t
;
using
walberla
::
uint_t
;
//////////////
// TYPEDEFS //
//////////////
using
CommunicationStencil_T
=
stencil
::
D3Q27
;
const
uint_t
FieldGhostLayers
=
4
;
using
ScalarField_T
=
field
::
GhostLayerField
<
real_t
,
1
>
;
using
VectorField_T
=
field
::
GhostLayerField
<
Vector3
<
real_t
>
,
1
>
;
using
MultiComponentField_T
=
field
::
GhostLayerField
<
real_t
,
3
>
;
/////////////////////
// BLOCK STRUCTURE //
/////////////////////
static
void
refinementSelection
(
SetupBlockForest
&
forest
,
const
uint_t
levels
)
{
const
AABB
&
domain
=
forest
.
getDomain
();
const
real_t
xSpan
=
domain
.
xSize
()
/
real_t
(
32
);
const
real_t
ySpan
=
domain
.
ySize
()
/
real_t
(
32
);
const
real_t
zSpan
=
domain
.
zSize
()
/
real_t
(
64
);
const
real_t
xMiddle
=
(
domain
.
xMin
()
+
domain
.
xMax
()
)
/
real_t
(
2
);
const
real_t
yMiddle
=
(
domain
.
yMin
()
+
domain
.
yMax
()
)
/
real_t
(
2
);
const
real_t
zMiddle
=
(
domain
.
zMin
()
+
domain
.
zMax
()
)
/
real_t
(
2
);
AABB
middleBox
(
xMiddle
-
xSpan
,
yMiddle
-
ySpan
,
zMiddle
+
zSpan
,
xMiddle
+
xSpan
,
yMiddle
+
ySpan
,
zMiddle
+
real_t
(
3
)
*
zSpan
);
AABB
shiftedBox
(
xMiddle
+
xSpan
,
yMiddle
+
ySpan
,
zMiddle
+
zSpan
,
xMiddle
+
real_t
(
3
)
*
xSpan
,
yMiddle
+
real_t
(
3
)
*
ySpan
,
zMiddle
+
real_t
(
3
)
*
zSpan
);
for
(
auto
block
=
forest
.
begin
();
block
!=
forest
.
end
();
++
block
)
{
if
(
block
->
getAABB
().
intersects
(
middleBox
)
||
block
->
getAABB
().
intersects
(
shiftedBox
)
)
if
(
block
->
getLevel
()
<
(
levels
-
uint_t
(
1
)
)
)
block
->
setMarker
(
true
);
}
}
static
void
workloadAndMemoryAssignment
(
SetupBlockForest
&
forest
)
{
for
(
auto
block
=
forest
.
begin
();
block
!=
forest
.
end
();
++
block
)
{
block
->
setWorkload
(
numeric_cast
<
workload_t
>
(
uint_t
(
1
)
<<
block
->
getLevel
()
)
);
block
->
setMemory
(
numeric_cast
<
memory_t
>
(
1
)
);
}
}
static
shared_ptr
<
StructuredBlockForest
>
createBlockStructure
(
const
uint_t
levels
,
const
uint_t
numberOfXBlocks
,
const
uint_t
numberOfYBlocks
,
const
uint_t
numberOfZBlocks
,
const
uint_t
numberOfXCellsPerBlock
,
const
uint_t
numberOfYCellsPerBlock
,
const
uint_t
numberOfZCellsPerBlock
,
const
bool
keepGlobalBlockInformation
=
false
)
{
// initialize SetupBlockForest = determine domain decomposition
SetupBlockForest
sforest
;
sforest
.
addRefinementSelectionFunction
(
std
::
bind
(
refinementSelection
,
std
::
placeholders
::
_1
,
levels
)
);
sforest
.
addWorkloadMemorySUIDAssignmentFunction
(
workloadAndMemoryAssignment
);
sforest
.
init
(
AABB
(
real_c
(
0
),
real_c
(
0
),
real_c
(
0
),
real_c
(
numberOfXBlocks
*
numberOfXCellsPerBlock
),
real_c
(
numberOfYBlocks
*
numberOfYCellsPerBlock
),
real_c
(
numberOfZBlocks
*
numberOfZCellsPerBlock
)
),
numberOfXBlocks
,
numberOfYBlocks
,
numberOfZBlocks
,
false
,
false
,
false
);
// calculate process distribution
const
memory_t
memoryLimit
=
math
::
Limits
<
memory_t
>::
inf
();
sforest
.
balanceLoad
(
blockforest
::
StaticLevelwiseCurveBalance
(
true
),
uint_c
(
MPIManager
::
instance
()
->
numProcesses
()
),
real_t
(
0
),
memoryLimit
,
true
);
MPIManager
::
instance
()
->
useWorldComm
();
// create StructuredBlockForest (encapsulates a newly created BlockForest)
shared_ptr
<
StructuredBlockForest
>
sbf
=
make_shared
<
StructuredBlockForest
>
(
make_shared
<
BlockForest
>
(
uint_c
(
MPIManager
::
instance
()
->
rank
()
),
sforest
,
keepGlobalBlockInformation
),
numberOfXCellsPerBlock
,
numberOfYCellsPerBlock
,
numberOfZCellsPerBlock
);
sbf
->
createCellBoundingBoxes
();
return
sbf
;
}
///////////////////////
// FIELD INITIALISER //
///////////////////////
template
<
typename
Field_T
>
void
clearField
(
std
::
shared_ptr
<
StructuredBlockForest
>
&
sbf
,
const
BlockDataID
&
fieldID
)
{
for
(
auto
it
=
sbf
->
begin
();
it
!=
sbf
->
end
();
++
it
)
{
auto
field
=
it
->
getData
<
Field_T
>
(
fieldID
);
WALBERLA_FOR_ALL_CELLS_INCLUDING_GHOST_LAYER_XYZ
(
field
,
for
(
uint_t
f
=
0
;
f
<
field
->
fSize
();
++
f
)
{
field
->
get
(
x
,
y
,
z
,
f
)
=
std
::
numeric_limits
<
typename
Field_T
::
value_type
>::
quiet_NaN
();
}
)
}
}
template
<
typename
Field_T
>
void
initialiseHomogeneously
(
std
::
shared_ptr
<
StructuredBlockForest
>
&
sbf
,
const
BlockDataID
&
fieldID
,
const
typename
Field_T
::
value_type
&
value
)
{
for
(
auto
blockIt
=
sbf
->
begin
();
blockIt
!=
sbf
->
end
();
++
blockIt
)
{
auto
field
=
blockIt
->
getData
<
Field_T
>
(
fieldID
);
WALBERLA_FOR_ALL_CELLS_XYZ
(
field
,
for
(
uint_t
f
=
0
;
f
<
field
->
fSize
();
++
f
)
{
field
->
get
(
x
,
y
,
z
,
f
)
=
value
;
}
)
}
}
template
<
typename
Field_T
>
void
initialiseCheckerboard
(
std
::
shared_ptr
<
StructuredBlockForest
>
&
sbf
,
const
BlockDataID
&
fieldID
,
const
typename
Field_T
::
value_type
&
value1
,
const
typename
Field_T
::
value_type
&
value2
)
{
for
(
auto
it
=
sbf
->
begin
();
it
!=
sbf
->
end
();
++
it
)
{
auto
field
=
it
->
getData
<
Field_T
>
(
fieldID
);
WALBERLA_FOR_ALL_CELLS_XYZ
(
field
,
typename
Field_T
::
value_type
value
;
((
x
+
y
+
z
)
%
2
==
0
)
?
value
=
value1
:
value
=
value2
;
for
(
uint_t
f
=
0
;
f
<
field
->
fSize
();
++
f
)
{
field
->
get
(
x
,
y
,
z
,
f
)
=
value
*
real_t
(
f
+
1
);
}
)
}
}
/////////////
// CHECKER //
/////////////
template
<
typename
CommunicationStencil_T
,
typename
Field_T
>
void
checkField
(
const
std
::
shared_ptr
<
StructuredBlockForest
>
&
sbf
,
const
BlockDataID
&
fieldID
)
{
for
(
auto
blockIt
=
sbf
->
begin
();
blockIt
!=
sbf
->
end
();
++
blockIt
)
{
Block
*
block
=
dynamic_cast
<
Block
*>
(
blockIt
.
get
());
// get fields
auto
*
field
=
blockIt
->
getData
<
Field_T
>
(
fieldID
);
for
(
auto
dir
=
CommunicationStencil_T
::
beginNoCenter
();
dir
!=
CommunicationStencil_T
::
end
();
++
dir
)
{
// get outer-most interior cells
CellInterval
ci
;
field
->
getSliceBeforeGhostLayer
(
*
dir
,
ci
);
// transform to global cell intervals
sbf
->
transformBlockLocalToGlobalCellInterval
(
ci
,
*
blockIt
);
const
auto
neighborIdx
=
blockforest
::
getBlockNeighborhoodSectionIndex
(
*
dir
);
Vector3
<
cell_idx_t
>
coarseShift
{
(
stencil
::
cx
[
*
dir
]
==
0
)
?
-
1
:
0
,
(
stencil
::
cy
[
*
dir
]
==
0
)
?
-
1
:
0
,
(
CommunicationStencil_T
::
D
==
3
)
?
((
stencil
::
cz
[
*
dir
]
==
0
)
?
-
1
:
0
)
:
0
};
auto
neighbourhoodSection
=
block
->
getNeighborhoodSection
(
neighborIdx
);
for
(
auto
nBlock
:
neighbourhoodSection
)
{
auto
neighbour
=
sbf
->
getBlock
(
nBlock
->
getId
());
auto
*
neighbourField
=
neighbour
->
template
getData
<
Field_T
>(
fieldID
);
for
(
auto
cell
=
ci
.
begin
();
cell
!=
ci
.
end
();
++
cell
)
{
// get physical coordinate
auto
cellCenter
=
sbf
->
getCellCenter
(
*
cell
,
sbf
->
getLevel
(
*
blockIt
));
auto
neighbourCell
=
sbf
->
getBlockLocalCell
(
*
neighbour
,
cellCenter
);
auto
localCell
=
sbf
->
getBlockLocalCell
(
*
blockIt
,
cellCenter
);
// check equal level communication -> must have the same values
if
(
block
->
neighborhoodSectionHasEquallySizedBlock
(
neighborIdx
))
{
for
(
uint_t
f
=
0
;
f
<
Field_T
::
F_SIZE
;
++
f
)
{
auto
fieldValue
=
field
->
get
(
localCell
,
f
);
auto
neighborValue
=
neighbourField
->
get
(
neighbourCell
,
f
);
WALBERLA_ASSERT_FLOAT_EQUAL
(
fieldValue
,
neighborValue
,
"Equal level communication failed."
)
}
}
// check fine-to-coarse communication
else
if
(
block
->
neighborhoodSectionHasLargerBlock
(
neighborIdx
))
{
Vector3
<
cell_idx_t
>
offset
{
(
localCell
.
x
()
%
2
==
0
)
?
+
1
:
-
1
,
(
localCell
.
y
()
%
2
==
0
)
?
+
1
:
-
1
,
(
localCell
.
z
()
%
2
==
0
)
?
+
1
:
-
1
};
for
(
uint_t
f
=
0
;
f
<
Field_T
::
F_SIZE
;
++
f
)
{
auto
fieldValue
=
field
->
get
(
localCell
[
0
]
,
localCell
[
1
]
,
localCell
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
+
offset
[
0
],
localCell
[
1
]
,
localCell
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
,
localCell
[
1
]
+
offset
[
1
],
localCell
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
+
offset
[
0
],
localCell
[
1
]
+
offset
[
1
],
localCell
[
2
],
f
);
if
(
CommunicationStencil_T
::
D
==
3
)
{
fieldValue
+=
field
->
get
(
localCell
[
0
]
,
localCell
[
1
]
,
localCell
[
2
]
+
offset
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
+
offset
[
0
],
localCell
[
1
]
,
localCell
[
2
]
+
offset
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
,
localCell
[
1
]
+
offset
[
1
],
localCell
[
2
]
+
offset
[
2
],
f
);
fieldValue
+=
field
->
get
(
localCell
[
0
]
+
offset
[
0
],
localCell
[
1
]
+
offset
[
1
],
localCell
[
2
]
+
offset
[
2
],
f
);
}
(
CommunicationStencil_T
::
D
==
3
)
?
fieldValue
/=
real_t
(
8
)
:
fieldValue
/=
real_t
(
4
);
auto
neighborValue
=
neighbourField
->
get
(
neighbourCell
,
f
);
WALBERLA_ASSERT_FLOAT_EQUAL
(
fieldValue
,
neighborValue
,
"Fine-to-coarse communication failed for scalar field."
)
}
}
// check coarse-to-fine communication
else
if
(
block
->
neighborhoodSectionHasSmallerBlocks
(
neighborIdx
))
{
// assure to not check outer most ghost layers of neighbour that are not affected by communication
CellInterval
neighbourGhostSlice
;
neighbourField
->
getGhostRegion
(
*
dir
,
neighbourGhostSlice
,
2
);
if
(
!
neighbourGhostSlice
.
contains
(
neighbourCell
))
continue
;
for
(
uint_t
f
=
0
;
f
<
Field_T
::
F_SIZE
;
++
f
)
{
auto
fieldValue
=
field
->
get
(
localCell
,
f
);
for
(
cell_idx_t
i
=
coarseShift
[
0
];
i
<=
0
;
++
i
)
{
for
(
cell_idx_t
j
=
coarseShift
[
1
];
j
<=
0
;
++
j
)
{
for
(
cell_idx_t
k
=
coarseShift
[
2
];
k
<=
0
;
++
k
)
{
auto
neighbourValue
=
neighbourField
->
get
(
neighbourCell
[
0
]
+
i
,
neighbourCell
[
1
]
+
j
,
neighbourCell
[
2
]
+
k
,
f
);
WALBERLA_ASSERT_FLOAT_EQUAL
(
fieldValue
,
neighbourValue
,
"Coarse-to-fine communication failed for global cell <"
<<
cell
->
x
()
<<
", "
<<
cell
->
y
()
<<
", "
<<
cell
->
z
()
<<
">."
)
}
}
}
}
}
else
{
WALBERLA_ABORT
(
"Something somewhere went terribly wrong!"
)
}
}
}
}
}
}
//////////
// MAIN //
//////////
int
main
(
int
argc
,
char
**
argv
)
{
debug
::
enterTestMode
();
mpi
::
Environment
env
(
argc
,
argv
);
logging
::
Logging
::
printHeaderOnStream
();
const
uint_t
levels
=
uint_t
(
4
);
const
uint_t
xBlocks
=
uint_t
(
4
);
const
uint_t
yBlocks
=
uint_t
(
4
);
const
uint_t
zBlocks
=
uint_t
(
4
);
const
uint_t
xCells
=
uint_t
(
10
);
const
uint_t
yCells
=
uint_t
(
10
);
const
uint_t
zCells
=
uint_t
(
10
);
auto
blocks
=
createBlockStructure
(
levels
,
xBlocks
,
yBlocks
,
zBlocks
,
xCells
,
yCells
,
zCells
);
// create fields
auto
scalarFieldID
=
field
::
addToStorage
<
ScalarField_T
>
(
blocks
,
"scalar field"
,
real_t
(
0
),
field
::
fzyx
,
FieldGhostLayers
);
auto
vectorFieldID
=
field
::
addToStorage
<
VectorField_T
>
(
blocks
,
"vector field"
,
Vector3
<
real_t
>
(
0
),
field
::
fzyx
,
FieldGhostLayers
);
auto
multiComponentFieldID
=
field
::
addToStorage
<
MultiComponentField_T
>
(
blocks
,
"multicomponent field"
,
real_t
(
0
),
field
::
fzyx
,
FieldGhostLayers
);
// create communication scheme
blockforest
::
communication
::
NonUniformBufferedScheme
<
CommunicationStencil_T
>
fieldCommunication
(
blocks
);
fieldCommunication
.
addPackInfo
(
std
::
make_shared
<
field
::
refinement
::
PackInfo
<
ScalarField_T
,
CommunicationStencil_T
>>
(
scalarFieldID
)
);
fieldCommunication
.
addPackInfo
(
std
::
make_shared
<
field
::
refinement
::
PackInfo
<
VectorField_T
,
CommunicationStencil_T
>>
(
vectorFieldID
)
);
fieldCommunication
.
addPackInfo
(
std
::
make_shared
<
field
::
refinement
::
PackInfo
<
MultiComponentField_T
,
CommunicationStencil_T
>>
(
multiComponentFieldID
)
);
///////////////////////////////
/// TEST HOMOGENEOUS FIELDS ///
///////////////////////////////
// clear fields
clearField
<
ScalarField_T
>
(
blocks
,
scalarFieldID
);
clearField
<
VectorField_T
>
(
blocks
,
vectorFieldID
);
clearField
<
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
);
// initialise fields
const
real_t
homogeneousValue
{
2
};
initialiseHomogeneously
<
ScalarField_T
>
(
blocks
,
scalarFieldID
,
homogeneousValue
);
initialiseHomogeneously
<
VectorField_T
>
(
blocks
,
vectorFieldID
,
Vector3
<
real_t
>
(
homogeneousValue
));
initialiseHomogeneously
<
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
,
homogeneousValue
);
// communicate
fieldCommunication
();
// check ghost layers
checkField
<
CommunicationStencil_T
,
ScalarField_T
>
(
blocks
,
scalarFieldID
);
checkField
<
CommunicationStencil_T
,
VectorField_T
>
(
blocks
,
vectorFieldID
);
checkField
<
CommunicationStencil_T
,
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
);
////////////////////////////////
/// TEST CHECKERBOARD FIELDS ///
////////////////////////////////
// clear fields
clearField
<
ScalarField_T
>
(
blocks
,
scalarFieldID
);
clearField
<
VectorField_T
>
(
blocks
,
vectorFieldID
);
clearField
<
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
);
// initialise fields
const
real_t
lowerValue
{
0
};
const
real_t
upperValue
{
2
};
initialiseCheckerboard
<
ScalarField_T
>
(
blocks
,
scalarFieldID
,
lowerValue
,
upperValue
);
initialiseCheckerboard
<
VectorField_T
>
(
blocks
,
vectorFieldID
,
Vector3
<
real_t
>
(
lowerValue
),
Vector3
<
real_t
>
(
upperValue
));
initialiseCheckerboard
<
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
,
lowerValue
,
upperValue
);
// communicate
fieldCommunication
();
// check ghost layers
checkField
<
CommunicationStencil_T
,
ScalarField_T
>
(
blocks
,
scalarFieldID
);
checkField
<
CommunicationStencil_T
,
VectorField_T
>
(
blocks
,
vectorFieldID
);
checkField
<
CommunicationStencil_T
,
MultiComponentField_T
>
(
blocks
,
multiComponentFieldID
);
logging
::
Logging
::
printFooterOnStream
();
return
EXIT_SUCCESS
;
}
}
// namespace nonuniform_buffered_scheme_test
int
main
(
int
argc
,
char
**
argv
)
{
return
nonuniform_buffered_scheme_test
::
main
(
argc
,
argv
);
}
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment