From 58d371d70bd4652c384fe81fee8a884012139870 Mon Sep 17 00:00:00 2001
From: Markus Holzer <markus.holzer@fau.de>
Date: Fri, 12 Feb 2021 18:01:07 +0100
Subject: [PATCH] Generated outflow bc
---
AUTHORS.txt | 1 +
.../UniformGridGPU/UniformGridGPU.py | 5 +-
.../UniformGridGenerated.py | 5 +-
apps/pythonmodule/PythonModule.cpp | 2 +
.../PhaseFieldAllenCahn/CPU/CMakeLists.txt | 3 +-
.../PhaseFieldAllenCahn/GPU/CMakeLists.txt | 3 +-
.../codegen/03_AdvancedLBMCodegen.py | 5 +-
python/lbmpy_walberla/__init__.py | 3 +-
.../lbmpy_walberla/additional_data_handler.py | 127 +++++++++++
python/lbmpy_walberla/boundary.py | 30 ++-
python/lbmpy_walberla/packinfo.py | 82 +++++++
.../lbmpy_walberla/walberla_lbm_generation.py | 1 -
.../additional_data_handler.py | 55 +++++
python/pystencils_walberla/boundary.py | 73 +++---
python/pystencils_walberla/codegen.py | 15 +-
python/pystencils_walberla/jinja_filters.py | 31 +++
.../templates/Boundary.tmpl.cpp | 22 +-
.../templates/Boundary.tmpl.h | 62 +++--
.../templates/Sweep.tmpl.cpp | 2 +-
src/lbm/field/QCriterion.h | 59 ++---
src/timeloop/SweepTimeloop.cpp | 10 +-
tests/lbm/CMakeLists.txt | 26 ++-
tests/lbm/codegen/CMakeLists.txt | 8 +
tests/lbm/codegen/GeneratedOutflowBC.cpp | 194 ++++++++++++++++
tests/lbm/codegen/GeneratedOutflowBC.prm | 24 ++
tests/lbm/codegen/GeneratedOutflowBC.py | 92 ++++++++
.../lbm/codegen/LbmPackInfoGenerationTest.py | 29 +++
tests/lbm/diff_packinfos.sh | 6 +
tests/timeloop/CMakeLists.txt | 4 +-
tests/timeloop/TimeloopAndSweepRegister.cpp | 212 +++++++-----------
30 files changed, 945 insertions(+), 246 deletions(-)
create mode 100644 python/lbmpy_walberla/additional_data_handler.py
create mode 100644 python/lbmpy_walberla/packinfo.py
create mode 100644 python/pystencils_walberla/additional_data_handler.py
create mode 100644 tests/lbm/codegen/CMakeLists.txt
create mode 100644 tests/lbm/codegen/GeneratedOutflowBC.cpp
create mode 100644 tests/lbm/codegen/GeneratedOutflowBC.prm
create mode 100644 tests/lbm/codegen/GeneratedOutflowBC.py
create mode 100644 tests/lbm/codegen/LbmPackInfoGenerationTest.py
create mode 100755 tests/lbm/diff_packinfos.sh
diff --git a/AUTHORS.txt b/AUTHORS.txt
index b0dba6103..2f956f1b1 100644
--- a/AUTHORS.txt
+++ b/AUTHORS.txt
@@ -13,6 +13,7 @@ Dominik Bartuschat
Ehsan Fattahi
Felix Winterhalter
Florian Schornbaum
+Frederik Hennig
Grigorii Drozdov
Helen Schottenhamml
Igor Ostanin
diff --git a/apps/benchmarks/UniformGridGPU/UniformGridGPU.py b/apps/benchmarks/UniformGridGPU/UniformGridGPU.py
index c1371278a..1a7973f6d 100644
--- a/apps/benchmarks/UniformGridGPU/UniformGridGPU.py
+++ b/apps/benchmarks/UniformGridGPU/UniformGridGPU.py
@@ -69,11 +69,10 @@ options_dict = {
'relaxation_rate': omega,
},
'cumulant': {
+ 'method': 'cumulant',
'stencil': 'D3Q19',
'compressible': True,
- 'method': 'mrt',
- 'cumulant': True,
- 'relaxation_rates': [0, omega, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
+ 'relaxation_rate': omega,
},
}
diff --git a/apps/benchmarks/UniformGridGenerated/UniformGridGenerated.py b/apps/benchmarks/UniformGridGenerated/UniformGridGenerated.py
index b9ca24ba1..fc587c445 100644
--- a/apps/benchmarks/UniformGridGenerated/UniformGridGenerated.py
+++ b/apps/benchmarks/UniformGridGenerated/UniformGridGenerated.py
@@ -55,11 +55,10 @@ options_dict = {
'relaxation_rate': omega,
},
'cumulant': {
+ 'method': 'cumulant',
'stencil': 'D3Q19',
'compressible': True,
- 'method': 'mrt',
- 'cumulant': True,
- 'relaxation_rates': [0, omega, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
+ 'relaxation_rate': omega,
},
}
diff --git a/apps/pythonmodule/PythonModule.cpp b/apps/pythonmodule/PythonModule.cpp
index b32f103dd..3059e3f05 100644
--- a/apps/pythonmodule/PythonModule.cpp
+++ b/apps/pythonmodule/PythonModule.cpp
@@ -40,11 +40,13 @@ using namespace walberla;
Field<walberla::real_t,1>,\
Field<walberla::real_t,2>,\
Field<walberla::real_t,3>,\
+ Field<walberla::real_t,4>,\
Field<walberla::real_t,9>,\
Field<walberla::real_t,15>,\
Field<walberla::real_t,19>,\
Field<walberla::real_t,27>,\
Field<walberla::int8_t,1>,\
+ Field<walberla::int32_t,1>,\
Field<walberla::int64_t,1>,\
Field<walberla::int64_t,2>,\
Field<walberla::int64_t,3>,\
diff --git a/apps/showcases/PhaseFieldAllenCahn/CPU/CMakeLists.txt b/apps/showcases/PhaseFieldAllenCahn/CPU/CMakeLists.txt
index 0e8f653f7..adacc3263 100644
--- a/apps/showcases/PhaseFieldAllenCahn/CPU/CMakeLists.txt
+++ b/apps/showcases/PhaseFieldAllenCahn/CPU/CMakeLists.txt
@@ -17,5 +17,4 @@ waLBerla_add_executable(NAME multiphaseCPU
FILES multiphase.cpp PythonExports.cpp InitializerFunctions.cpp contact.cpp CalculateNormals.cpp multiphase_codegen.py
DEPENDS blockforest core field postprocessing lbm geometry timeloop gui PhaseFieldCodeGenCPU)
-
-
+set_target_properties(multiphaseCPU PROPERTIES CXX_VISIBILITY_PRESET hidden)
diff --git a/apps/showcases/PhaseFieldAllenCahn/GPU/CMakeLists.txt b/apps/showcases/PhaseFieldAllenCahn/GPU/CMakeLists.txt
index c65a87749..51c692203 100644
--- a/apps/showcases/PhaseFieldAllenCahn/GPU/CMakeLists.txt
+++ b/apps/showcases/PhaseFieldAllenCahn/GPU/CMakeLists.txt
@@ -20,5 +20,4 @@ waLBerla_add_executable(NAME multiphaseGPU
FILES multiphase.cpp PythonExports.cpp InitializerFunctions.cpp CalculateNormals.cpp contact.cu multiphase_codegen.py
DEPENDS blockforest core cuda field postprocessing lbm geometry timeloop gui PhaseFieldCodeGenGPU)
-
-
+set_target_properties(multiphaseGPU PROPERTIES CXX_VISIBILITY_PRESET hidden)
diff --git a/apps/tutorials/codegen/03_AdvancedLBMCodegen.py b/apps/tutorials/codegen/03_AdvancedLBMCodegen.py
index a39790980..65a0602f2 100644
--- a/apps/tutorials/codegen/03_AdvancedLBMCodegen.py
+++ b/apps/tutorials/codegen/03_AdvancedLBMCodegen.py
@@ -36,9 +36,8 @@ optimization = {'cse_global': True,
# ==================
lbm_params = {'stencil': stencil,
- 'method': 'mrt_raw',
- 'relaxation_rates': [0, 0, 0, omega, omega, omega, 1, 1, 1],
- 'cumulant': True,
+ 'method': 'cumulant',
+ 'relaxation_rate': omega,
'compressible': True}
lbm_update_rule = create_lb_update_rule(optimization=optimization,
diff --git a/python/lbmpy_walberla/__init__.py b/python/lbmpy_walberla/__init__.py
index f7b022ab0..1a7b136a7 100644
--- a/python/lbmpy_walberla/__init__.py
+++ b/python/lbmpy_walberla/__init__.py
@@ -1,4 +1,5 @@
from .boundary import generate_boundary
from .walberla_lbm_generation import RefinementScaling, generate_lattice_model
+from .packinfo import generate_lb_pack_info
-__all__ = ['generate_lattice_model', 'RefinementScaling', 'generate_boundary']
+__all__ = ['generate_lattice_model', 'RefinementScaling', 'generate_boundary', 'generate_lb_pack_info']
diff --git a/python/lbmpy_walberla/additional_data_handler.py b/python/lbmpy_walberla/additional_data_handler.py
new file mode 100644
index 000000000..804c4953e
--- /dev/null
+++ b/python/lbmpy_walberla/additional_data_handler.py
@@ -0,0 +1,127 @@
+from pystencils.stencil import inverse_direction
+
+from lbmpy.advanced_streaming import AccessPdfValues, numeric_offsets, numeric_index
+from lbmpy.boundaries import ExtrapolationOutflow, UBB
+
+from pystencils_walberla.additional_data_handler import AdditionalDataHandler
+
+
+class UBBAdditionalDataHandler(AdditionalDataHandler):
+ def __init__(self, stencil, boundary_object):
+ assert isinstance(boundary_object, UBB)
+ self._boundary_object = boundary_object
+ super(UBBAdditionalDataHandler, self).__init__(stencil=stencil)
+
+ @property
+ def constructor_arguments(self):
+ return ", std::function<Vector3<real_t>(const Cell &, const shared_ptr<StructuredBlockForest>&, IBlock&)>& " \
+ "velocityCallback "
+
+ @property
+ def initialiser_list(self):
+ return "elementInitaliser(velocityCallback),"
+
+ @property
+ def additional_arguments_for_fill_function(self):
+ return "blocks, "
+
+ @property
+ def additional_parameters_for_fill_function(self):
+ return " const shared_ptr<StructuredBlockForest> &blocks, "
+
+ def data_initialisation(self, direction):
+ init_list = ["Vector3<real_t> InitialisatonAdditionalData = elementInitaliser(Cell(it.x(), it.y(), it.z()), "
+ "blocks, *block);", "element.vel_0 = InitialisatonAdditionalData[0];",
+ "element.vel_1 = InitialisatonAdditionalData[1];"]
+ if self._dim == 3:
+ init_list.append("element.vel_2 = InitialisatonAdditionalData[2];")
+
+ return "\n".join(init_list)
+
+ @property
+ def additional_member_variable(self):
+ return "std::function<Vector3<real_t>(const Cell &, const shared_ptr<StructuredBlockForest>&, IBlock&)> " \
+ "elementInitaliser; "
+
+
+class OutflowAdditionalDataHandler(AdditionalDataHandler):
+ def __init__(self, stencil, boundary_object, target='cpu', field_name='pdfs'):
+ assert isinstance(boundary_object, ExtrapolationOutflow)
+ self._boundary_object = boundary_object
+ self._stencil = boundary_object.stencil
+ self._lb_method = boundary_object.lb_method
+ self._normal_direction = boundary_object.normal_direction
+ self._field_name = field_name
+ self._target = target
+ super(OutflowAdditionalDataHandler, self).__init__(stencil=stencil)
+
+ assert sum([a != 0 for a in self._normal_direction]) == 1,\
+ "The outflow boundary is only implemented for straight walls at the moment."
+
+ @property
+ def constructor_arguments(self):
+ return f", BlockDataID {self._field_name}CPUID_" if self._target == 'gpu' else ""
+
+ @property
+ def initialiser_list(self):
+ return f"{self._field_name}CPUID({self._field_name}CPUID_)," if self._target == 'gpu' else ""
+
+ @property
+ def additional_field_data(self):
+ identifier = "CPU" if self._target == "gpu" else ""
+ return f"auto {self._field_name} = block->getData< field::GhostLayerField<double, " \
+ f"{len(self._stencil)}> >({self._field_name}{identifier}ID); "
+
+ def data_initialisation(self, direction_index):
+ pdf_acc = AccessPdfValues(self._boundary_object.stencil,
+ streaming_pattern=self._boundary_object.streaming_pattern,
+ timestep=self._boundary_object.zeroth_timestep,
+ streaming_dir='out')
+
+ init_list = []
+ for key, value in self.get_init_dict(pdf_acc, direction_index).items():
+ init_list.append(f"element.{key} = {self._field_name}->get({value});")
+
+ return "\n".join(init_list)
+
+ @property
+ def additional_member_variable(self):
+ return f"BlockDataID {self._field_name}CPUID;"
+
+ @property
+ def stencil_info(self):
+ stencil_info = []
+ for i, d in enumerate(self._stencil):
+ if any([a != 0 and b != 0 and a == b for a, b in zip(self._normal_direction, d)]):
+ direction = d if self._dim == 3 else d + (0,)
+ stencil_info.append((i, direction, ", ".join([str(e) for e in direction])))
+ return stencil_info
+
+ def get_init_dict(self, pdf_accessor, direction_index):
+ """The Extrapolation Outflow boundary needs additional data. This function provides a list of all values
+ which have to be initialised"""
+ position = ["it.x()", "it.y()", "it.z()"]
+ direction = self._stencil[direction_index]
+ inv_dir = self._stencil.index(inverse_direction(direction))
+
+ tangential_offset = tuple(offset - normal for offset, normal in zip(direction, self._normal_direction))
+
+ result = {}
+ pos = []
+ offsets = numeric_offsets(pdf_accessor.accs[inv_dir])
+ for p, o, t in zip(position, offsets, tangential_offset):
+ pos.append(p + f" + cell_idx_c({str(o + t)})")
+ if self._dim == 2:
+ pos.append("0")
+ pos.append(str(numeric_index(pdf_accessor.accs[inv_dir])[0]))
+ result[f'pdf'] = ', '.join(pos)
+
+ pos = []
+ for p, o, t in zip(position, offsets, tangential_offset):
+ pos.append(p + f" + cell_idx_c({str(o + t)})")
+ if self._dim == 2:
+ pos.append("0")
+ pos.append(str(numeric_index(pdf_accessor.accs[inv_dir])[0]))
+ result[f'pdf_nd'] = ', '.join(pos)
+
+ return result
diff --git a/python/lbmpy_walberla/boundary.py b/python/lbmpy_walberla/boundary.py
index 8f751f529..b0dafb7b3 100644
--- a/python/lbmpy_walberla/boundary.py
+++ b/python/lbmpy_walberla/boundary.py
@@ -1,5 +1,6 @@
import pystencils_walberla.boundary
from lbmpy.boundaries.boundaryhandling import create_lattice_boltzmann_boundary_kernel
+from lbmpy.advanced_streaming import Timestep, is_inplace
def generate_boundary(generation_context,
@@ -7,15 +8,29 @@ def generate_boundary(generation_context,
boundary_object,
lb_method,
field_name='pdfs',
+ streaming_pattern='pull',
+ always_generate_separate_classes=False,
+ additional_data_handler=None,
**create_kernel_params):
-
def boundary_creation_function(field, index_field, stencil, boundary_functor, target='cpu', openmp=True, **kwargs):
- return create_lattice_boltzmann_boundary_kernel(field,
- index_field,
- lb_method,
- boundary_functor,
- target=target,
- **kwargs)
+ pargs = (field, index_field, lb_method, boundary_functor)
+ kwargs = {'target': target, **kwargs}
+ if is_inplace(streaming_pattern) or always_generate_separate_classes:
+ return {
+ 'EvenSweep': create_lattice_boltzmann_boundary_kernel(*pargs,
+ streaming_pattern=streaming_pattern,
+ prev_timestep=Timestep.EVEN,
+ **kwargs),
+ 'OddSweep': create_lattice_boltzmann_boundary_kernel(*pargs,
+ streaming_pattern=streaming_pattern,
+ prev_timestep=Timestep.ODD,
+ **kwargs)
+ }
+ else:
+ return create_lattice_boltzmann_boundary_kernel(*pargs,
+ streaming_pattern=streaming_pattern,
+ prev_timestep=Timestep.BOTH,
+ **kwargs)
pystencils_walberla.boundary.generate_boundary(generation_context,
class_name,
@@ -25,4 +40,5 @@ def generate_boundary(generation_context,
index_shape=[len(lb_method.stencil)],
kernel_creation_function=boundary_creation_function,
namespace='lbm',
+ additional_data_handler=additional_data_handler,
**create_kernel_params)
diff --git a/python/lbmpy_walberla/packinfo.py b/python/lbmpy_walberla/packinfo.py
new file mode 100644
index 000000000..54a4e6755
--- /dev/null
+++ b/python/lbmpy_walberla/packinfo.py
@@ -0,0 +1,82 @@
+from collections import defaultdict
+from lbmpy.advanced_streaming.utility import Timestep, get_accessor, get_timesteps
+from lbmpy.advanced_streaming.communication import _extend_dir
+from pystencils.stencil import inverse_direction
+from pystencils_walberla.codegen import comm_directions, generate_pack_info
+from pystencils import Assignment, Field
+
+
+def generate_lb_pack_info(generation_context,
+ class_name_prefix: str,
+ stencil,
+ pdf_field,
+ streaming_pattern='pull',
+ lb_collision_rule=None,
+ always_generate_separate_classes=False,
+ namespace='lbm',
+ **create_kernel_params):
+ """Generates waLBerla MPI PackInfos for an LBM kernel, based on a given method
+ and streaming pattern. For in-place streaming patterns, two PackInfos are generated;
+ one for the even and another for the odd time steps.
+
+ Args:
+ generation_context: see documentation of `generate_sweep`
+ class_name_prefix: Prefix of the desired class name which will be extended with
+ 'Even' or 'Odd' for in-place kernels
+ stencil: The tuple of directions specifying the employed LB stencil.
+ pdf_field: pdf field for which the pack info is created
+ streaming_pattern: The employed streaming pattern.
+ lb_collision_rule: Optional. The collision rule defining the lattice boltzmann kernel, as returned
+ by `create_lb_collision_rule`. If specified, it will be scanned for non-local
+ accesses to other fields other than the PDF fields (as might be required for
+ computing gradients in coupled simulations), whose communication will then
+ be included in the PackInfo.
+ always_generate_separate_classes: If True, generate a pair of Even/Odd PackInfos even for a two-
+ fields kernel (i.e. the pull/push patterns). Otherwise, for two-fields
+ kernels, only one PackInfo class will be generated without a
+ suffix to its name.
+ namespace: inner namespace of the generated class
+ **create_kernel_params: remaining keyword arguments are passed to `pystencils.create_kernel`
+ """
+ timesteps = [Timestep.EVEN, Timestep.ODD] \
+ if always_generate_separate_classes \
+ else get_timesteps(streaming_pattern)
+
+ common_spec = defaultdict(set)
+
+ if lb_collision_rule is not None:
+ assignments = lb_collision_rule.all_assignments
+ reads = set()
+ for a in assignments:
+ if not isinstance(a, Assignment):
+ continue
+ reads.update(a.rhs.atoms(Field.Access))
+ for fa in reads:
+ assert all(abs(e) <= 1 for e in fa.offsets)
+ if all(offset == 0 for offset in fa.offsets):
+ continue
+ comm_direction = inverse_direction(fa.offsets)
+ for comm_dir in comm_directions(comm_direction):
+ common_spec[(comm_dir,)].add(fa.field.center(*fa.index))
+
+ for t in timesteps:
+ spec = common_spec.copy()
+ write_accesses = get_accessor(streaming_pattern, t).write(pdf_field, stencil)
+ for comm_dir in stencil:
+ if all(d == 0 for d in comm_dir):
+ continue
+
+ for streaming_dir in set(_extend_dir(comm_dir)) & set(stencil):
+ d = stencil.index(streaming_dir)
+ fa = write_accesses[d]
+ spec[(comm_dir,)].add(fa)
+
+ if t == Timestep.EVEN:
+ class_name_suffix = 'Even'
+ elif t == Timestep.ODD:
+ class_name_suffix = 'Odd'
+ else:
+ class_name_suffix = ''
+
+ class_name = class_name_prefix + class_name_suffix
+ generate_pack_info(generation_context, class_name, spec, namespace=namespace, **create_kernel_params)
diff --git a/python/lbmpy_walberla/walberla_lbm_generation.py b/python/lbmpy_walberla/walberla_lbm_generation.py
index a7563b76c..82867d8b8 100644
--- a/python/lbmpy_walberla/walberla_lbm_generation.py
+++ b/python/lbmpy_walberla/walberla_lbm_generation.py
@@ -6,7 +6,6 @@ from jinja2 import Environment, PackageLoader, StrictUndefined, Template
from sympy.tensor import IndexedBase
import pystencils as ps
-# from lbmpy.creationfunctions import create_lb_update_rule, update_with_default_parameters
from lbmpy.fieldaccess import CollideOnlyInplaceAccessor, StreamPullTwoFieldsAccessor
from lbmpy.relaxationrates import relaxation_rate_scaling
from lbmpy.stencils import get_stencil
diff --git a/python/pystencils_walberla/additional_data_handler.py b/python/pystencils_walberla/additional_data_handler.py
new file mode 100644
index 000000000..7abe7d5aa
--- /dev/null
+++ b/python/pystencils_walberla/additional_data_handler.py
@@ -0,0 +1,55 @@
+from pystencils.stencil import inverse_direction
+
+
+class AdditionalDataHandler:
+ """Base class that defines how to handle boundary conditions holding additional data."""
+
+ def __init__(self, stencil):
+ self._dim = len(stencil[0])
+
+ # waLBerla is a 3D framework. Therefore, a zero for the z index has to be added if we work in 2D
+ if self._dim == 2:
+ self._walberla_stencil = ()
+ for d in stencil:
+ d = d + (0,)
+ self._walberla_stencil = self._walberla_stencil + (d,)
+ else:
+ self._walberla_stencil = stencil
+
+ @property
+ def constructor_arguments(self):
+ return ""
+
+ @property
+ def initialiser_list(self):
+ return ""
+
+ @property
+ def additional_arguments_for_fill_function(self):
+ return ""
+
+ @property
+ def additional_parameters_for_fill_function(self):
+ return ""
+
+ @property
+ def additional_field_data(self):
+ return ""
+
+ def data_initialisation(self, direction_index):
+ return ""
+
+ @property
+ def additional_member_variable(self):
+ return ""
+
+ @property
+ def stencil_info(self):
+ return [(i, d, ", ".join([str(e) for e in d])) for i, d in enumerate(self._walberla_stencil)]
+
+ @property
+ def inverse_directions(self):
+ inv_dirs = []
+ for direction in self._walberla_stencil:
+ inv_dirs.append(self._walberla_stencil.index(inverse_direction(direction)))
+ return inv_dirs
diff --git a/python/pystencils_walberla/boundary.py b/python/pystencils_walberla/boundary.py
index 73d429ec5..0d251236c 100644
--- a/python/pystencils_walberla/boundary.py
+++ b/python/pystencils_walberla/boundary.py
@@ -1,3 +1,4 @@
+from collections import OrderedDict
import numpy as np
from jinja2 import Environment, PackageLoader, StrictUndefined
from pystencils import Field, FieldType
@@ -8,6 +9,7 @@ from pystencils.boundaries.createindexlist import (
from pystencils.data_types import TypedSymbol, create_type
from pystencils_walberla.codegen import KernelInfo, default_create_kernel_parameters
from pystencils_walberla.jinja_filters import add_pystencils_filters_to_jinja_env
+from pystencils_walberla.additional_data_handler import AdditionalDataHandler
def generate_boundary(generation_context,
@@ -20,7 +22,12 @@ def generate_boundary(generation_context,
kernel_creation_function=None,
target='cpu',
namespace='pystencils',
+ additional_data_handler=None,
**create_kernel_params):
+
+ if boundary_object.additional_data and additional_data_handler is None:
+ raise ValueError("Boundary object has additional data but you have not provided an AdditionalDataHandler.")
+
struct_name = "IndexInfo"
boundary_object.name = class_name
dim = len(neighbor_stencil[0])
@@ -44,36 +51,44 @@ def generate_boundary(generation_context,
if not kernel_creation_function:
kernel_creation_function = create_boundary_kernel
- kernel = kernel_creation_function(field, index_field, neighbor_stencil, boundary_object, **create_kernel_params)
- kernel.function_name = "boundary_" + boundary_object.name
- kernel.assumed_inner_stride_one = False
-
- # waLBerla is a 3D framework. Therefore, a zero for the z index has to be added if we work in 2D
- if dim == 2:
- stencil = ()
- for d in neighbor_stencil:
- d = d + (0,)
- stencil = stencil + (d,)
+ kernels = kernel_creation_function(field, index_field, neighbor_stencil, boundary_object, **create_kernel_params)
+ if isinstance(kernels, dict):
+ sweep_to_kernel_info_dict = OrderedDict()
+ dummy_kernel_info = None
+ for sweep_class, sweep_kernel in kernels.items():
+ sweep_kernel.function_name = "boundary_" + boundary_object.name + '_' + sweep_class
+ sweep_kernel.assumed_inner_stride_one = False
+ kernel_info = KernelInfo(sweep_kernel)
+ sweep_to_kernel_info_dict[sweep_class] = kernel_info
+ if dummy_kernel_info is None:
+ dummy_kernel_info = kernel_info
+ # elif not dummy_kernel_info.has_same_interface(kernel_info):
+ # raise ValueError("Multiple boundary sweeps must have the same kernel interface!")
+ multi_sweep = True
else:
- stencil = neighbor_stencil
+ multi_sweep = False
+ kernel = kernels
+ kernel.function_name = "boundary_" + boundary_object.name
+ kernel.assumed_inner_stride_one = False
+ kernel_info = KernelInfo(kernel)
+ sweep_to_kernel_info_dict = {'': kernel_info}
+ dummy_kernel_info = kernel_info
- stencil_info = [(i, d, ", ".join([str(e) for e in d])) for i, d in enumerate(stencil)]
- inv_dirs = []
- for direction in stencil:
- inverse_dir = tuple([-i for i in direction])
- inv_dirs.append(stencil.index(inverse_dir))
+ if additional_data_handler is None:
+ additional_data_handler = AdditionalDataHandler(stencil=neighbor_stencil)
context = {
'class_name': boundary_object.name,
+ 'sweep_classes': sweep_to_kernel_info_dict,
+ 'multi_sweep': multi_sweep,
+ 'dummy_kernel_info': dummy_kernel_info,
'StructName': struct_name,
'StructDeclaration': struct_from_numpy_dtype(struct_name, index_struct_dtype),
- 'kernel': KernelInfo(kernel),
- 'stencil_info': stencil_info,
- 'inverse_directions': inv_dirs,
'dim': dim,
'target': target,
'namespace': namespace,
- 'inner_or_boundary': boundary_object.inner_or_boundary
+ 'inner_or_boundary': boundary_object.inner_or_boundary,
+ 'additional_data_handler': additional_data_handler
}
env = Environment(loader=PackageLoader('pystencils_walberla'), undefined=StrictUndefined)
@@ -83,8 +98,8 @@ def generate_boundary(generation_context,
source = env.get_template('Boundary.tmpl.cpp').render(**context)
source_extension = "cpp" if target == "cpu" else "cu"
- generation_context.write_file("{}.h".format(class_name), header)
- generation_context.write_file("{}.{}".format(class_name, source_extension), source)
+ generation_context.write_file(f"{class_name}.h", header)
+ generation_context.write_file(f"{class_name}.{source_extension}", source)
def generate_staggered_boundary(generation_context, class_name, boundary_object,
@@ -102,23 +117,23 @@ def generate_staggered_flux_boundary(generation_context, class_name, boundary_ob
def struct_from_numpy_dtype(struct_name, numpy_dtype):
- result = "struct %s { \n" % (struct_name,)
+ result = f"struct {struct_name} {{ \n"
equality_compare = []
constructor_params = []
constructor_initializer_list = []
for name, (sub_type, offset) in numpy_dtype.fields.items():
pystencils_type = create_type(sub_type)
- result += " %s %s;\n" % (pystencils_type, name)
+ result += f" {pystencils_type} {name};\n"
if name in boundary_index_array_coordinate_names or name == direction_member_name:
- constructor_params.append("%s %s_" % (pystencils_type, name))
- constructor_initializer_list.append("%s(%s_)" % (name, name))
+ constructor_params.append(f"{pystencils_type} {name}_")
+ constructor_initializer_list.append(f"{name}({name}_)")
else:
- constructor_initializer_list.append("%s()" % name)
+ constructor_initializer_list.append(f"{name}()")
if pystencils_type.is_float():
- equality_compare.append("floatIsEqual(%s, o.%s)" % (name, name))
+ equality_compare.append(f"floatIsEqual({name}, o.{name})")
else:
- equality_compare.append("%s == o.%s" % (name, name))
+ equality_compare.append(f"{name} == o.{name}")
result += " %s(%s) : %s {}\n" % \
(struct_name, ", ".join(constructor_params), ", ".join(constructor_initializer_list))
diff --git a/python/pystencils_walberla/codegen.py b/python/pystencils_walberla/codegen.py
index 37013cba6..896a795cc 100644
--- a/python/pystencils_walberla/codegen.py
+++ b/python/pystencils_walberla/codegen.py
@@ -20,7 +20,7 @@ __all__ = ['generate_sweep', 'generate_pack_info', 'generate_pack_info_for_field
def generate_sweep(generation_context, class_name, assignments,
namespace='pystencils', field_swaps=(), staggered=False, varying_parameters=(),
- inner_outer_split=False,
+ inner_outer_split=False, ghost_layers_to_include=0,
**create_kernel_params):
"""Generates a waLBerla sweep from a pystencils representation.
@@ -44,6 +44,8 @@ def generate_sweep(generation_context, class_name, assignments,
the C++ class constructor even if the kernel does not need them.
inner_outer_split: if True generate a sweep that supports separate iteration over inner and outer regions
to allow for communication hiding.
+ ghost_layers_to_include: determines how many ghost layers should be included for the Sweep.
+ This is relevant if a setter kernel should also set correct values to the ghost layers.
**create_kernel_params: remaining keyword arguments are passed to `pystencils.create_kernel`
"""
create_kernel_params = default_create_kernel_parameters(generation_context, create_kernel_params)
@@ -78,6 +80,7 @@ def generate_sweep(generation_context, class_name, assignments,
'class_name': class_name,
'target': create_kernel_params.get("target", "cpu"),
'headers': get_headers(ast),
+ 'ghost_layers_to_include': ghost_layers_to_include
}
header = env.get_template("Sweep.tmpl.h").render(**jinja_context)
source = env.get_template("Sweep.tmpl.cpp").render(**jinja_context)
@@ -93,6 +96,7 @@ def generate_sweep(generation_context, class_name, assignments,
'target': create_kernel_params.get("target", "cpu"),
'field': representative_field,
'headers': get_headers(ast),
+ 'ghost_layers_to_include': 0
}
header = env.get_template("SweepInnerOuter.tmpl.h").render(**jinja_context)
source = env.get_template("SweepInnerOuter.tmpl.cpp").render(**jinja_context)
@@ -132,7 +136,7 @@ def generate_pack_info_from_kernel(generation_context, class_name: str, assignme
class_name: name of the generated class
assignments: list of assignments from the compute kernel - generates PackInfo for "pull" part only
i.e. the kernel is expected to only write to the center
- kind:
+ kind: can either be pull or push
**create_kernel_params: remaining keyword arguments are passed to `pystencils.create_kernel`
"""
assert kind in ('push', 'pull')
@@ -240,7 +244,6 @@ def generate_pack_info(generation_context, class_name: str,
pack_kernels[direction_strings] = KernelInfo(pack_ast)
unpack_kernels[direction_strings] = KernelInfo(unpack_ast)
elements_per_cell[direction_strings] = len(terms)
-
fused_kernel = create_kernel([Assignment(buffer.center, t) for t in all_accesses], **create_kernel_params)
fused_kernel.assumed_inner_stride_one = create_kernel_params['cpu_vectorize_info']['assume_inner_stride_one']
@@ -333,6 +336,12 @@ class KernelInfo:
self.varying_parameters = tuple(varying_parameters)
self.parameters = ast.get_parameters() # cache parameters here
+ def has_same_interface(self, other):
+ return self.temporary_fields == other.temporary_fields \
+ and self.field_swaps == other.field_swaps \
+ and self.varying_parameters == other.varying_parameters \
+ and self.parameters == other.parameters
+
def get_vectorize_instruction_set(generation_context):
if generation_context.optimize_for_localhost:
diff --git a/python/pystencils_walberla/jinja_filters.py b/python/pystencils_walberla/jinja_filters.py
index 60903be9e..7158f8168 100644
--- a/python/pystencils_walberla/jinja_filters.py
+++ b/python/pystencils_walberla/jinja_filters.py
@@ -338,6 +338,26 @@ def generate_constructor_parameters(kernel_info, parameters_to_ignore=None):
return ", ".join(parameter_list + varying_parameters)
+def generate_constructor_call_arguments(kernel_info, parameters_to_ignore=None):
+ if parameters_to_ignore is None:
+ parameters_to_ignore = []
+
+ varying_parameters = []
+ if hasattr(kernel_info, 'varying_parameters'):
+ varying_parameters = kernel_info.varying_parameters
+ varying_parameter_names = tuple(e[1] for e in varying_parameters)
+ parameters_to_ignore += kernel_info.temporary_fields + varying_parameter_names
+
+ parameter_list = []
+ for param in kernel_info.parameters:
+ if param.is_field_pointer and param.field_name not in parameters_to_ignore:
+ parameter_list.append("%sID" % (param.field_name, ))
+ elif not param.is_field_parameter and param.symbol.name not in parameters_to_ignore:
+ parameter_list.append(f'{param.symbol.name}_')
+ varying_parameters = ["%s_" % e for e in varying_parameter_names]
+ return ", ".join(parameter_list + varying_parameters)
+
+
@jinja2.contextfilter
def generate_members(ctx, kernel_info, parameters_to_ignore=(), only_fields=False):
ast = kernel_info.ast
@@ -382,14 +402,25 @@ def generate_destructor(kernel_info, class_name):
return temporary_constructor.format(contents=contents, class_name=class_name)
+@jinja2.contextfilter
+def nested_class_method_definition_prefix(ctx, nested_class_name):
+ outer_class = ctx['class_name']
+ if len(nested_class_name) == 0:
+ return outer_class
+ else:
+ return outer_class + '::' + nested_class_name
+
+
def add_pystencils_filters_to_jinja_env(jinja_env):
jinja_env.filters['generate_definition'] = generate_definition
jinja_env.filters['generate_declaration'] = generate_declaration
jinja_env.filters['generate_members'] = generate_members
jinja_env.filters['generate_constructor_parameters'] = generate_constructor_parameters
jinja_env.filters['generate_constructor_initializer_list'] = generate_constructor_initializer_list
+ jinja_env.filters['generate_constructor_call_arguments'] = generate_constructor_call_arguments
jinja_env.filters['generate_call'] = generate_call
jinja_env.filters['generate_block_data_to_field_extraction'] = generate_block_data_to_field_extraction
jinja_env.filters['generate_swaps'] = generate_swaps
jinja_env.filters['generate_refs_for_kernel_parameters'] = generate_refs_for_kernel_parameters
jinja_env.filters['generate_destructor'] = generate_destructor
+ jinja_env.filters['nested_class_method_definition_prefix'] = nested_class_method_definition_prefix
diff --git a/python/pystencils_walberla/templates/Boundary.tmpl.cpp b/python/pystencils_walberla/templates/Boundary.tmpl.cpp
index 33f799107..ef1ec3b7b 100644
--- a/python/pystencils_walberla/templates/Boundary.tmpl.cpp
+++ b/python/pystencils_walberla/templates/Boundary.tmpl.cpp
@@ -51,7 +51,11 @@ namespace {{namespace}} {
#endif
-{{kernel|generate_definition}}
+{% for sweep_class, sweep_kernel in sweep_classes.items() %}
+
+{{sweep_kernel|generate_definition}}
+
+{% endfor %}
#ifdef __GNUC__
#pragma GCC diagnostic pop
@@ -61,8 +65,10 @@ namespace {{namespace}} {
#pragma pop
#endif
+{% for sweep_class, sweep_kernel in sweep_classes.items() %}
+
-void {{class_name}}::run( IBlock * block, IndexVectors::Type type {% if target == 'gpu'%}, cudaStream_t stream {%endif%})
+void {{sweep_class|nested_class_method_definition_prefix}}::run( IBlock * block, IndexVectors::Type type {% if target == 'gpu'%}, cudaStream_t stream {%endif%})
{
auto * indexVectors = block->getData<IndexVectors>(indexVectorID);
int64_t indexVectorSize = int64_c( indexVectors->indexVector(type).size() );
@@ -77,25 +83,27 @@ void {{class_name}}::run( IBlock * block, IndexVectors::Type type {% if target =
uint8_t * _data_indexVector = reinterpret_cast<uint8_t*>(pointer);
- {{kernel|generate_block_data_to_field_extraction(['indexVector', 'indexVectorSize'])|indent(4)}}
- {{kernel|generate_call(spatial_shape_symbols=['indexVectorSize'], stream='stream')|indent(4)}}
+ {{sweep_kernel|generate_block_data_to_field_extraction(['indexVector', 'indexVectorSize'])|indent(4)}}
+ {{sweep_kernel|generate_refs_for_kernel_parameters(prefix='', parameters_to_ignore=['indexVectorSize'], ignore_fields=True)|indent(4) }}
+ {{sweep_kernel|generate_call(spatial_shape_symbols=['indexVectorSize'], stream='stream')|indent(4)}}
}
-void {{class_name}}::operator() ( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
+void {{sweep_class|nested_class_method_definition_prefix}}::operator() ( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
{
run( block, IndexVectors::ALL{% if target == 'gpu'%}, stream {%endif%});
}
-void {{class_name}}::inner( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
+void {{sweep_class|nested_class_method_definition_prefix}}::inner( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
{
run( block, IndexVectors::INNER{% if target == 'gpu'%}, stream {%endif%} );
}
-void {{class_name}}::outer( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
+void {{sweep_class|nested_class_method_definition_prefix}}::outer( IBlock * block{% if target == 'gpu'%}, cudaStream_t stream {%endif%} )
{
run( block, IndexVectors::OUTER{% if target == 'gpu'%}, stream {%endif%} );
}
+{% endfor %}
} // namespace {{namespace}}
} // namespace walberla
diff --git a/python/pystencils_walberla/templates/Boundary.tmpl.h b/python/pystencils_walberla/templates/Boundary.tmpl.h
index c856adcea..c0a39c32b 100644
--- a/python/pystencils_walberla/templates/Boundary.tmpl.h
+++ b/python/pystencils_walberla/templates/Boundary.tmpl.h
@@ -17,13 +17,13 @@
//! \\author pystencils
//======================================================================================================================
-
#include "core/DataTypes.h"
{% if target is equalto 'cpu' -%}
#include "field/GhostLayerField.h"
{%- elif target is equalto 'gpu' -%}
#include "cuda/GPUField.h"
+#include "cuda/FieldCopy.h"
{%- endif %}
#include "domain_decomposition/BlockDataID.h"
#include "domain_decomposition/IBlock.h"
@@ -64,7 +64,7 @@ public:
NUM_TYPES = 3
};
- IndexVectors() : cpuVectors_(NUM_TYPES) {}
+ IndexVectors() = default;
bool operator==(IndexVectors & other) { return other.cpuVectors_ == cpuVectors_; }
{% if target == 'gpu' -%}
@@ -72,14 +72,14 @@ public:
for( auto & gpuVec: gpuVectors_)
cudaFree( gpuVec );
}
- {% endif %}
+ {% endif -%}
CpuIndexVector & indexVector(Type t) { return cpuVectors_[t]; }
{{StructName}} * pointerCpu(Type t) { return &(cpuVectors_[t][0]); }
{% if target == 'gpu' -%}
{{StructName}} * pointerGpu(Type t) { return gpuVectors_[t]; }
- {% endif %}
+ {% endif -%}
void syncGPU()
{
@@ -100,39 +100,66 @@ public:
}
private:
- std::vector<CpuIndexVector> cpuVectors_;
+ std::vector<CpuIndexVector> cpuVectors_{NUM_TYPES};
{% if target == 'gpu' -%}
using GpuIndexVector = {{StructName}} *;
std::vector<GpuIndexVector> gpuVectors_;
- {% endif %}
+ {%- endif %}
+ };
+
+ {% if multi_sweep %}
+ {% for sweep_class_name, sweep_kernel_info in sweep_classes.items() %}
+ class {{sweep_class_name}}
+ {
+ public:
+ {{sweep_class_name}} ( {{sweep_kernel_info|generate_constructor_parameters(['indexVectorSize'])}} )
+ : {{ sweep_kernel_info|generate_constructor_initializer_list(['indexVectorSize']) }} {};
+
+ void operator() ( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+ void inner( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+ void outer( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+
+ private:
+ void run( IBlock * block, IndexVectors::Type type{% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+
+ {{sweep_kernel_info|generate_members(['indexVectorSize'])|indent(12)}}
};
+ {{sweep_class_name}} get{{sweep_class_name}} ()
+ {
+ return {{sweep_class_name}} ( {{sweep_kernel_info|generate_constructor_call_arguments(['indexVectorSize'])|indent(12)}} );
+ }
+ {% endfor %}
+ {% endif %}
{{class_name}}( const shared_ptr<StructuredBlockForest> & blocks,
- {{kernel|generate_constructor_parameters(['indexVector', 'indexVectorSize'])}} )
- : {{ kernel|generate_constructor_initializer_list(['indexVector', 'indexVectorSize']) }}
+ {{dummy_kernel_info|generate_constructor_parameters(['indexVector', 'indexVectorSize'])}}{{additional_data_handler.constructor_arguments}})
+ :{{additional_data_handler.initialiser_list}} {{ dummy_kernel_info|generate_constructor_initializer_list(['indexVector', 'indexVectorSize']) }}
{
auto createIdxVector = []( IBlock * const , StructuredBlockStorage * const ) { return new IndexVectors(); };
indexVectorID = blocks->addStructuredBlockData< IndexVectors >( createIdxVector, "IndexField_{{class_name}}");
};
+ {% if not multi_sweep %}
+
void operator() ( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
void inner( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
void outer( IBlock * block {% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+ {% endif %}
template<typename FlagField_T>
void fillFromFlagField( const shared_ptr<StructuredBlockForest> & blocks, ConstBlockDataID flagFieldID,
FlagUID boundaryFlagUID, FlagUID domainFlagUID)
{
for( auto blockIt = blocks->begin(); blockIt != blocks->end(); ++blockIt )
- fillFromFlagField<FlagField_T>( &*blockIt, flagFieldID, boundaryFlagUID, domainFlagUID );
+ fillFromFlagField<FlagField_T>({{additional_data_handler.additional_arguments_for_fill_function}}&*blockIt, flagFieldID, boundaryFlagUID, domainFlagUID );
}
template<typename FlagField_T>
- void fillFromFlagField( IBlock * block, ConstBlockDataID flagFieldID,
+ void fillFromFlagField({{additional_data_handler.additional_parameters_for_fill_function}}IBlock * block, ConstBlockDataID flagFieldID,
FlagUID boundaryFlagUID, FlagUID domainFlagUID )
{
auto * indexVectors = block->getData< IndexVectors > ( indexVectorID );
@@ -140,8 +167,8 @@ public:
auto & indexVectorInner = indexVectors->indexVector(IndexVectors::INNER);
auto & indexVectorOuter = indexVectors->indexVector(IndexVectors::OUTER);
-
auto * flagField = block->getData< FlagField_T > ( flagFieldID );
+ {{additional_data_handler.additional_field_data|indent(4)}}
if( !(flagField->flagExists(boundaryFlagUID) && flagField->flagExists(domainFlagUID) ))
return;
@@ -152,7 +179,6 @@ public:
auto inner = flagField->xyzSize();
inner.expand( cell_idx_t(-1) );
-
indexVectorAll.clear();
indexVectorInner.clear();
indexVectorOuter.clear();
@@ -161,15 +187,15 @@ public:
{
if( ! isFlagSet(it, domainFlag) )
continue;
-
- {%- for dirIdx, dirVec, offset in stencil_info %}
+ {%- for dirIdx, dirVec, offset in additional_data_handler.stencil_info %}
if ( isFlagSet( it.neighbor({{offset}} {%if dim == 3%}, 0 {%endif %}), boundaryFlag ) )
{
{% if inner_or_boundary -%}
auto element = {{StructName}}(it.x(), it.y(), {%if dim == 3%} it.z(), {%endif %} {{dirIdx}} );
{% else -%}
- auto element = {{StructName}}(it.x() + cell_idx_c({{dirVec[0]}}), it.y() + cell_idx_c({{dirVec[1]}}), {%if dim == 3%} it.z() + cell_idx_c({{dirVec[2]}}), {%endif %} {{inverse_directions[dirIdx]}} );
+ auto element = {{StructName}}(it.x() + cell_idx_c({{dirVec[0]}}), it.y() + cell_idx_c({{dirVec[1]}}), {%if dim == 3%} it.z() + cell_idx_c({{dirVec[2]}}), {%endif %} {{additional_data_handler.inverse_directions[dirIdx]}} );
{% endif -%}
+ {{additional_data_handler.data_initialisation(dirIdx)|indent(16)}}
indexVectorAll.push_back( element );
if( inner.contains( it.x(), it.y(), it.z() ) )
indexVectorInner.push_back( element );
@@ -178,16 +204,18 @@ public:
}
{% endfor %}
}
-
indexVectors->syncGPU();
}
private:
+ {% if not multi_sweep %}
void run( IBlock * block, IndexVectors::Type type{% if target == 'gpu'%}, cudaStream_t stream = 0 {%endif%});
+ {% endif %}
BlockDataID indexVectorID;
+ {{additional_data_handler.additional_member_variable|indent(4)}}
public:
- {{kernel|generate_members(('indexVector', 'indexVectorSize'))|indent(4)}}
+ {{dummy_kernel_info|generate_members(('indexVector', 'indexVectorSize'))|indent(4)}}
};
diff --git a/python/pystencils_walberla/templates/Sweep.tmpl.cpp b/python/pystencils_walberla/templates/Sweep.tmpl.cpp
index b3b12c6b3..b26d9c6db 100644
--- a/python/pystencils_walberla/templates/Sweep.tmpl.cpp
+++ b/python/pystencils_walberla/templates/Sweep.tmpl.cpp
@@ -59,7 +59,7 @@ void {{class_name}}::operator()( IBlock * block{%if target is equalto 'gpu'%} ,
{
{{kernel|generate_block_data_to_field_extraction|indent(4)}}
{{kernel|generate_refs_for_kernel_parameters(prefix='this->', ignore_fields=True)|indent(4) }}
- {{kernel|generate_call(stream='stream')|indent(4)}}
+ {{kernel|generate_call(ghost_layers_to_include=ghost_layers_to_include, stream='stream')|indent(4)}}
{{kernel|generate_swaps|indent(4)}}
}
diff --git a/src/lbm/field/QCriterion.h b/src/lbm/field/QCriterion.h
index eff9b8015..c68111778 100644
--- a/src/lbm/field/QCriterion.h
+++ b/src/lbm/field/QCriterion.h
@@ -28,36 +28,40 @@
// You should never use these functions directly, always refer to the member functions
// of PdfField or the free functions that can be found in MacroscopicValueCalculation.h
-namespace walberla {
-namespace lbm {
-
-struct QCriterion {
+namespace walberla
+{
+namespace lbm
+{
+struct QCriterion
+{
template< typename VelocityField_T, typename Filter_T >
- static inline real_t get(const VelocityField_T & velocityField, const Filter_T & filter,
- const cell_idx_t x, const cell_idx_t y, const cell_idx_t z,
- real_t dx = real_t(1), real_t dy = real_t(1), real_t dz = real_t(1)) {
+ static inline real_t get(const VelocityField_T& velocityField, const Filter_T& filter, const cell_idx_t x,
+ const cell_idx_t y, const cell_idx_t z, real_t dx = real_t(1), real_t dy = real_t(1),
+ real_t dz = real_t(1))
+ {
const auto one = cell_idx_t(1);
- if(filter(x,y,z) && filter(x+one,y,z) && filter(x-one,y,z) && filter(x,y+one,z)
- && filter(x,y-one,z) && filter(x,y,z+one) && filter(x,y,z-one)) {
- const Vector3<real_t> xa = velocityField.get(x+one,y,z);
- const Vector3<real_t> xb = velocityField.get(x-one,y,z);
- const Vector3<real_t> ya = velocityField.get(x,y+one,z);
- const Vector3<real_t> yb = velocityField.get(x,y-one,z);
- const Vector3<real_t> za = velocityField.get(x,y,z+one);
- const Vector3<real_t> zb = velocityField.get(x,y,z-one);
+ auto f(velocityField.flattenedShallowCopy());
+
+ if (filter(x, y, z) && filter(x + one, y, z) && filter(x - one, y, z) && filter(x, y + one, z) &&
+ filter(x, y - one, z) && filter(x, y, z + one) && filter(x, y, z - one))
+ {
+ Vector3< real_t > xa(f->get(x + one, y, z, 0), f->get(x + one, y, z, 1), f->get(x + one, y, z, 2));
+ Vector3< real_t > xb(f->get(x - one, y, z, 0), f->get(x - one, y, z, 1), f->get(x - one, y, z, 2));
+ Vector3< real_t > ya(f->get(x, y + one, z, 0), f->get(x, y + one, z, 1), f->get(x, y + one, z, 2));
+ Vector3< real_t > yb(f->get(x, y - one, z, 0), f->get(x, y - one, z, 1), f->get(x, y - one, z, 2));
+ Vector3< real_t > za(f->get(x, y, z + one, 0), f->get(x, y, z + one, 1), f->get(x, y, z + one, 2));
+ Vector3< real_t > zb(f->get(x, y, z - one, 0), f->get(x, y, z - one, 1), f->get(x, y, z - one, 2));
return calculate(xa, xb, ya, yb, za, zb, dx, dy, dz);
}
-
return real_t(0);
}
- static inline real_t calculate(const Vector3<real_t> xa, const Vector3<real_t> xb,
- const Vector3<real_t> ya, const Vector3<real_t> yb,
- const Vector3<real_t> za, const Vector3<real_t> zb,
- const real_t dx, const real_t dy, const real_t dz) {
-
+ static inline real_t calculate(const Vector3< real_t > xa, const Vector3< real_t > xb, const Vector3< real_t > ya,
+ const Vector3< real_t > yb, const Vector3< real_t > za, const Vector3< real_t > zb,
+ const real_t dx, const real_t dy, const real_t dz)
+ {
const auto halfInvDx = real_t(0.5) / dx;
const auto halfInvDy = real_t(0.5) / dy;
const auto halfInvDz = real_t(0.5) / dz;
@@ -75,18 +79,17 @@ struct QCriterion {
const real_t duzdz = (za[2] - zb[2]) * halfInvDz;
// Q = 1/2 * (||W||² - ||S||²)
- real_t sNormSq = duxdx*duxdx + duydy*duydy + duzdz*duzdz +
- real_t(0.5)*(duxdy+duydx)*(duxdy+duydx) + real_t(0.5)*(duydz+duzdy)*(duydz+duzdy) +
- real_t(0.5)*(duxdz+duzdx)*(duxdz+duzdx);
+ real_t sNormSq = duxdx * duxdx + duydy * duydy + duzdz * duzdz + real_t(0.5) * (duxdy + duydx) * (duxdy + duydx) +
+ real_t(0.5) * (duydz + duzdy) * (duydz + duzdy) +
+ real_t(0.5) * (duxdz + duzdx) * (duxdz + duzdx);
- real_t omegaNormSq = real_t(0.5)*(duxdz-duzdx)*(duxdz-duzdx) +
- real_t(0.5)*(duxdy-duydx)*(duxdy-duydx) +
- real_t(0.5)*(duydz-duzdy)*(duydz-duzdy);
+ real_t omegaNormSq = real_t(0.5) * (duxdz - duzdx) * (duxdz - duzdx) +
+ real_t(0.5) * (duxdy - duydx) * (duxdy - duydx) +
+ real_t(0.5) * (duydz - duzdy) * (duydz - duzdy);
return real_t(0.5) * (omegaNormSq - sNormSq);
}
};
-
} // namespace lbm
} // namespace walberla
diff --git a/src/timeloop/SweepTimeloop.cpp b/src/timeloop/SweepTimeloop.cpp
index 7fc72a38a..15b407191 100644
--- a/src/timeloop/SweepTimeloop.cpp
+++ b/src/timeloop/SweepTimeloop.cpp
@@ -54,10 +54,8 @@ void SweepTimeloop::doTimeStep(const Set<SUID> &selectors)
}
Sweep * selectedSweep = s.sweep.getUnique( selectors + bi->getState() );
- if( !selectedSweep )
- WALBERLA_ABORT("Selecting Sweep " << sweepIt->first << ": " <<
- "Ambiguous, or no sweep selected. Check your selector " <<
- selectors + bi->getState() << std::endl << s.sweep);
+ if (!selectedSweep)
+ continue;
WALBERLA_LOG_PROGRESS_SECTION()
{
@@ -111,9 +109,7 @@ void SweepTimeloop::doTimeStep(const Set<SUID> &selectors, WcTimingPool &timing)
Sweep * selectedSweep = s.sweep.getUnique( selectors + bi->getState(), sweepName );
if( !selectedSweep )
- WALBERLA_ABORT("Selecting Sweep " << sweepIt->first << ": " <<
- "Ambiguous, or no sweep selected. Check your selector " <<
- selectors + bi->getState() << std::endl << s.sweep);
+ continue;
WALBERLA_LOG_PROGRESS("Running sweep \"" << sweepName << "\" on block " << bi->getId() );
diff --git a/tests/lbm/CMakeLists.txt b/tests/lbm/CMakeLists.txt
index b17f65cfa..a2ddd6690 100644
--- a/tests/lbm/CMakeLists.txt
+++ b/tests/lbm/CMakeLists.txt
@@ -76,6 +76,21 @@ waLBerla_execute_test( NAME QCriterionTest )
# Code Generation
if( WALBERLA_BUILD_WITH_CODEGEN )
+add_subdirectory(codegen)
+
+waLBerla_generate_target_from_python(NAME GeneratedOutflowBCGenerated
+ FILE codegen/GeneratedOutflowBC.py
+ OUT_FILES GeneratedOutflowBC_Sweep.cpp GeneratedOutflowBC_Sweep.h
+ GeneratedOutflowBC_MacroSetter.cpp GeneratedOutflowBC_MacroSetter.h
+ GeneratedOutflowBC_Dynamic_UBB.cpp GeneratedOutflowBC_Dynamic_UBB.h
+ GeneratedOutflowBC_Static_UBB.cpp GeneratedOutflowBC_Static_UBB.h
+ GeneratedOutflowBC_NoSlip.cpp GeneratedOutflowBC_NoSlip.h
+ GeneratedOutflowBC_Outflow.cpp GeneratedOutflowBC_Outflow.h
+ GeneratedOutflowBC_PackInfo.cpp GeneratedOutflowBC_PackInfo.h
+ GeneratedOutflowBC_InfoHeader.h)
+waLBerla_compile_test( FILES codegen/GeneratedOutflowBC.cpp DEPENDS GeneratedOutflowBCGenerated)
+waLBerla_execute_test( NAME GeneratedOutflowBC COMMAND $<TARGET_FILE:GeneratedOutflowBC> ${CMAKE_CURRENT_SOURCE_DIR}/codegen/GeneratedOutflowBC.prm )
+
waLBerla_generate_target_from_python(NAME LbCodeGenerationExampleGenerated
FILE codegen/LbCodeGenerationExample.py
@@ -95,4 +110,13 @@ waLBerla_generate_target_from_python(NAME FieldLayoutAndVectorizationTestGenerat
FieldLayoutAndVectorizationTest_ZYXF_NoVec_LatticeModel.cpp FieldLayoutAndVectorizationTest_ZYXF_NoVec_LatticeModel.h )
waLBerla_compile_test( FILES codegen/FieldLayoutAndVectorizationTest.cpp DEPENDS FieldLayoutAndVectorizationTestGenerated)
-endif()
\ No newline at end of file
+waLBerla_generate_target_from_python(NAME LbmPackInfoGenerationTestCodegen FILE codegen/LbmPackInfoGenerationTest.py
+ OUT_FILES AccessorBasedPackInfoEven.cpp AccessorBasedPackInfoEven.h
+ AccessorBasedPackInfoOdd.cpp AccessorBasedPackInfoOdd.h
+ FromKernelPackInfoPull.cpp FromKernelPackInfoPull.h
+ FromKernelPackInfoPush.cpp FromKernelPackInfoPush.h)
+
+waLBerla_link_files_to_builddir( "diff_packinfos.sh" )
+waLBerla_execute_test( NAME LbmPackInfoGenerationDiffTest COMMAND bash diff_packinfos.sh )
+
+endif()
diff --git a/tests/lbm/codegen/CMakeLists.txt b/tests/lbm/codegen/CMakeLists.txt
new file mode 100644
index 000000000..b149eda3f
--- /dev/null
+++ b/tests/lbm/codegen/CMakeLists.txt
@@ -0,0 +1,8 @@
+#############################################################################################################################
+#
+# Tests for lbm module
+#
+#############################################################################################################################
+
+waLBerla_link_files_to_builddir( "*.prm" )
+waLBerla_link_files_to_builddir( "ChannelFlowCodeGenParameter.py" )
\ No newline at end of file
diff --git a/tests/lbm/codegen/GeneratedOutflowBC.cpp b/tests/lbm/codegen/GeneratedOutflowBC.cpp
new file mode 100644
index 000000000..66311c7eb
--- /dev/null
+++ b/tests/lbm/codegen/GeneratedOutflowBC.cpp
@@ -0,0 +1,194 @@
+//======================================================================================================================
+//
+// 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 GeneratedOutflowBC.cpp
+//! \author Markus Holzer <markus.holzer@fau.de>
+//! \brief Shear flow with dynamic UBB on the left (W) with linear flow profile from zero to u_max (flow in x-direction)
+//! along the y-direction. The upper wall (N) is a static UBB which applies u_max in x-direction.
+//! On the right (E) an outflow boundary is used and the lower wall (S) is a No-Slip wall.
+//! It is tested if the shear flow is correctly propagated through the entire domain after n timesteps.
+//
+//======================================================================================================================
+#include "blockforest/Initialization.h"
+#include "blockforest/communication/UniformBufferedScheme.h"
+
+#include "core/Environment.h"
+#include "core/timing/RemainingTimeLogger.h"
+
+#include "field/AddToStorage.h"
+#include "field/vtk/VTKWriter.h"
+
+#include "geometry/InitBoundaryHandling.h"
+#include "timeloop/SweepTimeloop.h"
+
+// Generated Files
+#include "GeneratedOutflowBC_Dynamic_UBB.h"
+#include "GeneratedOutflowBC_InfoHeader.h"
+#include "GeneratedOutflowBC_MacroSetter.h"
+#include "GeneratedOutflowBC_NoSlip.h"
+#include "GeneratedOutflowBC_Outflow.h"
+#include "GeneratedOutflowBC_PackInfo.h"
+#include "GeneratedOutflowBC_Static_UBB.h"
+#include "GeneratedOutflowBC_Sweep.h"
+
+using namespace walberla;
+
+using PackInfo_T = lbm::GeneratedOutflowBC_PackInfo;
+using flag_t = walberla::uint8_t;
+using FlagField_T = FlagField< flag_t >;
+
+auto pdfFieldAdder = [](IBlock* const block, StructuredBlockStorage* const storage) {
+ return new PdfField_T(storage->getNumberOfXCells(*block), storage->getNumberOfYCells(*block),
+ storage->getNumberOfZCells(*block), uint_t(1), field::fzyx,
+ make_shared< field::AllocateAligned< real_t, 64 > >());
+};
+
+////////////////////////////////////////////
+// Linear Velocity Profile for left wall //
+//////////////////////////////////////////
+
+class ShearProfile
+{
+ public:
+
+ ShearProfile( real_t inflow_velocity ) :
+ inflow_velocity_( inflow_velocity ) {}
+
+ Vector3< real_t > operator()( const Cell& pos, const shared_ptr< StructuredBlockForest >& SbF, IBlock& block ) const;
+
+ private:
+
+ const real_t inflow_velocity_;
+}; // class ShearProfile
+
+Vector3< real_t > ShearProfile::operator()( const Cell& pos, const shared_ptr< StructuredBlockForest >& SbF, IBlock& block ) const
+{
+ Cell globalCell;
+ CellInterval domain = SbF->getDomainCellBB();
+ real_t h_y = domain.yMax() - domain.yMin();
+ SbF->transformBlockLocalToGlobalCell(globalCell, block, pos);
+
+ real_t u = inflow_velocity_ * (globalCell[1] / h_y);
+
+ Vector3< real_t > result(u, 0.0, 0.0);
+ return result;
+}
+
+//////////
+// MAIN //
+//////////
+
+int main(int argc, char** argv)
+{
+ walberla::Environment walberlaEnv(argc, argv);
+
+ auto blocks = blockforest::createUniformBlockGridFromConfig(walberlaEnv.config());
+
+ // read parameters
+ auto parameters = walberlaEnv.config()->getOneBlock("Parameters");
+
+ const real_t omega = parameters.getParameter< real_t >("omega", real_c(1.4));
+ const real_t u_max = parameters.getParameter< real_t >("u_max", real_t(0.05));
+ const uint_t timesteps = parameters.getParameter< uint_t >("timesteps", uint_c(10));
+
+ const double remainingTimeLoggerFrequency =
+ parameters.getParameter< double >("remainingTimeLoggerFrequency", 3.0); // in seconds
+
+ // create fields
+ BlockDataID pdfFieldID = blocks->addStructuredBlockData< PdfField_T >(pdfFieldAdder, "PDFs");
+ BlockDataID velFieldID = field::addToStorage< VelocityField_T >(blocks, "velocity", real_t(0), field::fzyx);
+ BlockDataID densityFieldID = field::addToStorage< ScalarField_T >(blocks, "density", real_t(0), field::fzyx);
+
+ BlockDataID flagFieldId = field::addFlagFieldToStorage< FlagField_T >(blocks, "flag field");
+
+ pystencils::GeneratedOutflowBC_MacroSetter setterSweep(pdfFieldID, velFieldID);
+ for (auto& block : *blocks)
+ setterSweep(&block);
+
+ // create and initialize boundary handling
+ const FlagUID fluidFlagUID("Fluid");
+
+ auto boundariesConfig = walberlaEnv.config()->getOneBlock("Boundaries");
+
+ ShearProfile velocityCallback{u_max};
+ std::function< Vector3< real_t >(const Cell&, const shared_ptr< StructuredBlockForest >&, IBlock&) >
+ velocity_initialisation = velocityCallback;
+
+ lbm::GeneratedOutflowBC_Dynamic_UBB ubb_dynamic(blocks, pdfFieldID, velocity_initialisation);
+ lbm::GeneratedOutflowBC_Static_UBB ubb_static(blocks, pdfFieldID, u_max);
+ lbm::GeneratedOutflowBC_NoSlip noSlip(blocks, pdfFieldID);
+ lbm::GeneratedOutflowBC_Outflow outflow(blocks, pdfFieldID);
+
+ geometry::initBoundaryHandling< FlagField_T >(*blocks, flagFieldId, boundariesConfig);
+ geometry::setNonBoundaryCellsToDomain< FlagField_T >(*blocks, flagFieldId, fluidFlagUID);
+
+ ubb_dynamic.fillFromFlagField< FlagField_T >(blocks, flagFieldId, FlagUID("UBB_Inflow"), fluidFlagUID);
+ ubb_static.fillFromFlagField< FlagField_T >(blocks, flagFieldId, FlagUID("UBB_Wall"), fluidFlagUID);
+ noSlip.fillFromFlagField< FlagField_T >(blocks, flagFieldId, FlagUID("NoSlip"), fluidFlagUID);
+ outflow.fillFromFlagField< FlagField_T >(blocks, flagFieldId, FlagUID("Outflow"), fluidFlagUID);
+
+ // create time loop
+ SweepTimeloop timeloop(blocks->getBlockStorage(), timesteps);
+
+ // create communication for PdfField
+ blockforest::communication::UniformBufferedScheme< Stencil_T > communication(blocks);
+ communication.addPackInfo(make_shared< PackInfo_T >(pdfFieldID));
+
+ pystencils::GeneratedOutflowBC_Sweep UpdateSweep(densityFieldID, pdfFieldID, velFieldID, omega);
+
+ // add LBM sweep and communication to time loop
+ timeloop.add() << BeforeFunction(communication, "communication") << Sweep(noSlip, "noSlip boundary");
+ timeloop.add() << Sweep(ubb_dynamic, "ubb inflow");
+ timeloop.add() << Sweep(ubb_static, "ubb wall");
+ timeloop.add() << Sweep(outflow, "outflow boundary");
+ timeloop.add() << Sweep(UpdateSweep, "LB stream & collide");
+
+ // log remaining time
+ timeloop.addFuncAfterTimeStep(timing::RemainingTimeLogger(timeloop.getNrOfTimeSteps(), remainingTimeLoggerFrequency),
+ "remaining time logger");
+
+ // VTK Writer
+ uint_t vtkWriteFrequency = parameters.getParameter< uint_t >("vtkWriteFrequency", 0);
+ if (vtkWriteFrequency > 0)
+ {
+ auto vtkOutput = vtk::createVTKOutput_BlockData(*blocks, "GeneratedOutflowBC_VTK", vtkWriteFrequency, 0, false,
+ "vtk_out", "simulation_step", false, true, true, false, 0);
+
+ auto velWriter = make_shared< field::VTKWriter< VelocityField_T > >(velFieldID, "velocity");
+ auto densityWriter = make_shared< field::VTKWriter< ScalarField_T > >(densityFieldID, "density");
+
+ vtkOutput->addCellDataWriter(velWriter);
+ vtkOutput->addCellDataWriter(densityWriter);
+
+ timeloop.addFuncBeforeTimeStep(vtk::writeFiles(vtkOutput), "VTK Output");
+ }
+ timeloop.run();
+
+ CellInterval domain = blocks->getDomainCellBB();
+ real_t h_y = domain.yMax() - domain.yMin();
+ for (auto& block : *blocks)
+ {
+ auto velField = block.getData<VelocityField_T>(velFieldID);
+ WALBERLA_FOR_ALL_CELLS_XYZ
+ (
+ velField,
+ Cell globalCell;
+ blocks->transformBlockLocalToGlobalCell(globalCell, block, Cell(x, y, z));
+ WALBERLA_CHECK_FLOAT_EQUAL_EPSILON(velField->get(x, y, z, 0), u_max * (globalCell[1] / h_y), 0.01)
+ )
+ }
+
+ return EXIT_SUCCESS;
+}
diff --git a/tests/lbm/codegen/GeneratedOutflowBC.prm b/tests/lbm/codegen/GeneratedOutflowBC.prm
new file mode 100644
index 000000000..b53043d9b
--- /dev/null
+++ b/tests/lbm/codegen/GeneratedOutflowBC.prm
@@ -0,0 +1,24 @@
+
+Parameters
+{
+ omega 1.8;
+ timesteps 500;
+ u_max 0.05;
+ vtkWriteFrequency 0;
+ remainingTimeLoggerFrequency 3; // in seconds
+}
+
+DomainSetup
+{
+ blocks < 1, 1, 1 >;
+ cellsPerBlock < 30, 30, 1 >;
+ periodic < 0, 0, 0 >;
+}
+
+Boundaries
+{
+ Border { direction W; walldistance -1; flag UBB_Inflow; }
+ Border { direction E; walldistance -1; flag Outflow; }
+ Border { direction S; walldistance -1; flag NoSlip; }
+ Border { direction N; walldistance -1; flag UBB_Wall; }
+}
diff --git a/tests/lbm/codegen/GeneratedOutflowBC.py b/tests/lbm/codegen/GeneratedOutflowBC.py
new file mode 100644
index 000000000..59a916fec
--- /dev/null
+++ b/tests/lbm/codegen/GeneratedOutflowBC.py
@@ -0,0 +1,92 @@
+from pystencils.field import fields
+from lbmpy.macroscopic_value_kernels import macroscopic_values_setter
+from lbmpy.stencils import get_stencil
+from lbmpy.creationfunctions import create_lb_method, create_lb_update_rule
+from lbmpy.boundaries import NoSlip, UBB, ExtrapolationOutflow
+from lbmpy_walberla.additional_data_handler import UBBAdditionalDataHandler, OutflowAdditionalDataHandler
+from pystencils_walberla import CodeGeneration, generate_sweep
+from lbmpy_walberla import RefinementScaling, generate_boundary, generate_lb_pack_info
+
+import sympy as sp
+
+stencil = get_stencil("D2Q9")
+q = len(stencil)
+dim = len(stencil[0])
+
+pdfs, pdfs_tmp = fields(f"pdfs({q}), pdfs_tmp({q}): double[{dim}D]", layout='fzyx')
+velocity_field, density_field = fields(f"velocity({dim}), density(1) : double[{dim}D]", layout='fzyx')
+omega = sp.Symbol("omega")
+u_max = sp.Symbol("u_max")
+
+output = {
+ 'density': density_field,
+ 'velocity': velocity_field
+}
+
+options = {'method': 'cumulant',
+ 'stencil': stencil,
+ 'relaxation_rate': omega,
+ 'galilean_correction': len(stencil) == 27,
+ 'field_name': 'pdfs',
+ 'output': output,
+ 'optimization': {'symbolic_field': pdfs,
+ 'symbolic_temporary_field': pdfs_tmp,
+ 'cse_global': False,
+ 'cse_pdfs': False}}
+
+method = create_lb_method(**options)
+
+# getter & setter
+setter_assignments = macroscopic_values_setter(method, velocity=velocity_field.center_vector,
+ pdfs=pdfs, density=1)
+
+# opt = {'instruction_set': 'sse', 'assume_aligned': True, 'nontemporal': False, 'assume_inner_stride_one': True}
+
+update_rule = create_lb_update_rule(lb_method=method, **options)
+
+info_header = f"""
+using namespace walberla;
+#include "stencil/D{dim}Q{q}.h"
+using Stencil_T = walberla::stencil::D{dim}Q{q};
+using PdfField_T = GhostLayerField<real_t, {q}>;
+using VelocityField_T = GhostLayerField<real_t, {dim}>;
+using ScalarField_T = GhostLayerField<real_t, 1>;
+ """
+
+stencil = method.stencil
+
+with CodeGeneration() as ctx:
+ # sweeps
+ generate_sweep(ctx, 'GeneratedOutflowBC_Sweep', update_rule, field_swaps=[(pdfs, pdfs_tmp)])
+ generate_sweep(ctx, 'GeneratedOutflowBC_MacroSetter', setter_assignments)
+
+ # boundaries
+ ubb_dynamic = UBB(lambda *args: None, dim=dim)
+ ubb_data_handler = UBBAdditionalDataHandler(stencil, ubb_dynamic)
+
+ if dim == 2:
+ ubb_static = UBB([sp.Symbol("u_max"), 0])
+ else:
+ ubb_static = UBB([sp.Symbol("u_max"), 0, 0])
+
+ outflow = ExtrapolationOutflow(stencil[4], method)
+ outflow_data_handler = OutflowAdditionalDataHandler(stencil, outflow)
+
+ # Dynamic UBB which is used to produce a specific velocity profile at the inflow.
+ # Note that the additional data handler is needed for that kind of boundary.
+ generate_boundary(ctx, 'GeneratedOutflowBC_Dynamic_UBB', ubb_dynamic, method,
+ additional_data_handler=ubb_data_handler)
+
+ # Static UBB which is used to apply a certain velocity u_max at the upper wall in x-direction
+ generate_boundary(ctx, 'GeneratedOutflowBC_Static_UBB', ubb_static, method)
+
+ generate_boundary(ctx, 'GeneratedOutflowBC_NoSlip', NoSlip(), method)
+
+ generate_boundary(ctx, 'GeneratedOutflowBC_Outflow', outflow, method,
+ additional_data_handler=outflow_data_handler)
+
+ # communication
+ generate_lb_pack_info(ctx, 'GeneratedOutflowBC_PackInfo', stencil, pdfs)
+
+ # Info header containing correct template definitions for stencil and field
+ ctx.write_file("GeneratedOutflowBC_InfoHeader.h", info_header)
diff --git a/tests/lbm/codegen/LbmPackInfoGenerationTest.py b/tests/lbm/codegen/LbmPackInfoGenerationTest.py
new file mode 100644
index 000000000..5fab9c7ab
--- /dev/null
+++ b/tests/lbm/codegen/LbmPackInfoGenerationTest.py
@@ -0,0 +1,29 @@
+from lbmpy.creationfunctions import create_lb_collision_rule, create_lb_update_rule
+from lbmpy.advanced_streaming import Timestep
+from lbmpy.stencils import get_stencil
+from pystencils_walberla import CodeGeneration, generate_pack_info_from_kernel
+from lbmpy_walberla.packinfo import generate_lb_pack_info
+from pystencils.field import Field
+
+with CodeGeneration() as ctx:
+ streaming_pattern = 'aa'
+ target = 'cpu'
+ stencil = get_stencil('D3Q19')
+ dim = len(stencil[0])
+ values_per_cell = len(stencil)
+ collision_rule = create_lb_collision_rule(method='srt', stencil=stencil)
+ pdf_field = Field.create_generic('pdfs', dim, index_shape=(values_per_cell,), layout='fzyx')
+ optimization = {
+ 'symbolic_field': pdf_field,
+ 'target': target
+ }
+
+ # Generate PackInfo specifically for streaming pattern
+ generate_lb_pack_info(ctx, 'AccessorBasedPackInfo', stencil, pdf_field,
+ streaming_pattern=streaming_pattern, target=target, namespace='pystencils')
+
+ # Generate reference using the alternating pull/push approach
+ update_rule_odd = create_lb_update_rule(collision_rule=collision_rule, optimization=optimization,
+ streaming_pattern=streaming_pattern, timestep=Timestep.ODD)
+ generate_pack_info_from_kernel(ctx, 'FromKernelPackInfoPull', update_rule_odd, kind='pull', target=target)
+ generate_pack_info_from_kernel(ctx, 'FromKernelPackInfoPush', update_rule_odd, kind='push', target=target)
diff --git a/tests/lbm/diff_packinfos.sh b/tests/lbm/diff_packinfos.sh
new file mode 100755
index 000000000..bfa89c5ef
--- /dev/null
+++ b/tests/lbm/diff_packinfos.sh
@@ -0,0 +1,6 @@
+#!/bin/bash
+
+REGEX='^((#include)|(void)|(uint_t))'
+cd default_codegen
+diff -u -B <(grep -vP "$REGEX" FromKernelPackInfoPull.cpp) <(grep -vP "$REGEX" AccessorBasedPackInfoEven.cpp) || exit 1
+diff -u -B <(grep -vP "$REGEX" FromKernelPackInfoPush.cpp) <(grep -vP "$REGEX" AccessorBasedPackInfoOdd.cpp) || exit 1
diff --git a/tests/timeloop/CMakeLists.txt b/tests/timeloop/CMakeLists.txt
index 265f266fd..862c3b648 100644
--- a/tests/timeloop/CMakeLists.txt
+++ b/tests/timeloop/CMakeLists.txt
@@ -5,5 +5,5 @@
###################################################################################################
-#waLBerla_compile_test( FILES TimeloopAndSweepRegister.cpp DEPENDS field blockforest )
-#waLBerla_execute_test(NAME TimeloopAndSweepRegister )
+waLBerla_compile_test( FILES TimeloopAndSweepRegister.cpp DEPENDS field blockforest )
+waLBerla_execute_test(NAME TimeloopAndSweepRegister )
diff --git a/tests/timeloop/TimeloopAndSweepRegister.cpp b/tests/timeloop/TimeloopAndSweepRegister.cpp
index 028e5fad7..944210821 100644
--- a/tests/timeloop/TimeloopAndSweepRegister.cpp
+++ b/tests/timeloop/TimeloopAndSweepRegister.cpp
@@ -1,175 +1,129 @@
//======================================================================================================================
//
-// This file is part of waLBerla. waLBerla is free software: you can
+// 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
+// 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
+//
+// 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 TimeloopAndSweepRegister.cpp
//! \ingroup timeloop
-//! \author Martin Bauer <martin.bauer@fau.de>
+//! \author Markus Holzer <markus.holzer@fau.de>
//! \brief test cases that test the registering of Sweeps at timeloop
//
//======================================================================================================================
-#include "timeloop/SweepTimeloop.h"
+#include "blockforest/Initialization.h"
+
#include "core/DataTypes.h"
+#include "core/Environment.h"
#include "core/debug/TestSubsystem.h"
-#include "core/logging/Logging.h"
+
+#include "field/Field.h"
+
+#include "timeloop/SweepTimeloop.h"
#include <string>
#include <vector>
-
-using namespace std;
using namespace walberla;
+using Field_T = Field< uint_t, 1 >;
+
+auto FieldAdder = [](IBlock* const block, StructuredBlockStorage* const storage) {
+ return new Field_T(storage->getNumberOfXCells(*block), storage->getNumberOfYCells(*block),
+ storage->getNumberOfZCells(*block), uint_t(0.0), field::fzyx,
+ make_shared< field::AllocateAligned< uint_t, 64 > >());
+};
-class GeneralSweep
+class Sweep1
{
- public:
- GeneralSweep(const string & name, vector<string> & vec)
- : myName(name), outVec(vec)
- {}
+ public:
+ Sweep1(BlockDataID fieldID) : fieldID_(fieldID) {}
+ void operator()(IBlock* block)
+ {
+ auto field = block->getData< Field_T >(fieldID_);
- void operator() (IBlock * ){
- outVec.push_back(myName);
- }
+ for (auto iter = field->begin(); iter != field->end(); ++iter)
+ *iter += 1;
+ }
- private:
- string myName;
- vector<string> & outVec;
+ private:
+ BlockDataID fieldID_;
};
-
-class GeneralFunction
+class Sweep2
{
- public:
- GeneralFunction(const string & name, vector<string> & vec)
- : myName(name), outVec(vec)
- {}
+ public:
+ Sweep2(BlockDataID fieldID) : fieldID_(fieldID) {}
- void operator() (){
- outVec.push_back(myName);
- }
-
- private:
- string myName;
- vector<string> & outVec;
-};
+ void operator()(IBlock* block)
+ {
+ auto field = block->getData< Field_T >(fieldID_);
+ for (auto iter = field->begin(); iter != field->end(); ++iter)
+ *iter += 2;
+ }
+ private:
+ BlockDataID fieldID_;
+};
-int main()
+int main(int argc, char** argv)
{
debug::enterTestMode();
+ mpi::Environment env(argc, argv);
- vector<string> expectedSequence;
- vector<string> sequence;
-
- SUID cpuSelect("CPU");
- SUID gpuSelect("GPU");
-
- SUID srtSelect("SRT");
- SUID mrtSelect("MRT");
-
-
- //FIXME put a real test in here
- shared_ptr<SweepTimeloop> tl = make_shared<SweepTimeloop>(shared_ptr<BlockStorage>(),100);
-
- typedef SweepTimeloop::SweepHandle SH;
-
- SH sweep1 = tl->addSweep( GeneralSweep("CPU1",sequence), cpuSelect);
- tl->addSweep(sweep1, GeneralSweep("GPU1",sequence), gpuSelect);
-
- SH sweep2 = tl->addSweep( GeneralSweep("CPU2",sequence), cpuSelect);
- tl->addSweep(sweep2, GeneralSweep("GPU2",sequence), gpuSelect);
+ std::vector<std::string> expectedSequence;
+ std::vector<std::string> sequence;
+ SUID sweepSelect1("Sweep1");
+ SUID sweepSelect2("Sweep2");
- tl->addFuncBeforeSweep(sweep1,GeneralFunction("Pre1",sequence));
- tl->addFuncAfterSweep (sweep1,GeneralFunction("Post1",sequence));
+ shared_ptr< StructuredBlockForest > blocks = blockforest::createUniformBlockGrid(
+ uint_c(4), uint_c(2), uint_c(2), uint_c(10), uint_c(10), uint_c(10), real_c(1), false, false, false, false);
- tl->addFuncBeforeSweep(sweep2,GeneralFunction("Pre2",sequence));
- tl->addFuncAfterSweep (sweep2,GeneralFunction("Post2",sequence));
+ BlockDataID fieldID = blocks->addStructuredBlockData< Field_T >(FieldAdder, "Test Field");
- typedef Timeloop::FctHandle FH;
- FH preTs = tl->addFuncBeforeTimeStep( GeneralFunction("PreTimestepCPU",sequence),cpuSelect,srtSelect);
- tl->addFuncBeforeTimeStep(preTs, GeneralFunction("PreTimestepGPU",sequence),gpuSelect,srtSelect);
- FH postTs = tl->addFuncAfterTimeStep ( GeneralFunction("PostTimestepCPU",sequence),cpuSelect,mrtSelect);
- tl->addFuncAfterTimeStep (postTs,GeneralFunction("PostTimestepGPU",sequence),gpuSelect,mrtSelect);
+ for (auto& block : *blocks)
+ {
+ if (block.getAABB().min()[0] < 20)
+ block.setState(sweepSelect1);
+ else
+ block.setState(sweepSelect2);
+ }
+ uint_t timesteps = 10;
+ SweepTimeloop timeloop(blocks->getBlockStorage(), timesteps);
- //---------- First Run - CPU Selector ---------------------------------------------
+ timeloop.add() << Sweep(Sweep1(fieldID), "Sweep 1", sweepSelect1, sweepSelect2);
+ timeloop.add() << Sweep(Sweep2(fieldID), "Sweep 2", sweepSelect2, sweepSelect1);
- expectedSequence.push_back("PreTimestepCPU");
- expectedSequence.push_back("Pre1");
- expectedSequence.push_back("CPU1");
- expectedSequence.push_back("Post1");
- expectedSequence.push_back("Pre2");
- expectedSequence.push_back("CPU2");
- expectedSequence.push_back("Post2");
- expectedSequence.push_back("PostTimestepCPU");
+ WcTimingPool timingPool;
- tl->singleStep(cpuSelect);
-
- WALBERLA_CHECK_EQUAL(expectedSequence.size(), sequence.size());
- WALBERLA_CHECK( equal(expectedSequence.begin(),expectedSequence.end(), sequence.begin() ) );
-
- expectedSequence.clear();
- sequence.clear();
-
-
- // ------------ Second Run - GPU Selector -------------------------------------------
-
- expectedSequence.push_back("PreTimestepGPU");
- expectedSequence.push_back("Pre1");
- expectedSequence.push_back("GPU1");
- expectedSequence.push_back("Post1");
- expectedSequence.push_back("Pre2");
- expectedSequence.push_back("GPU2");
- expectedSequence.push_back("Post2");
- expectedSequence.push_back("PostTimestepGPU");
-
- tl->singleStep(gpuSelect);
-
- WALBERLA_CHECK_EQUAL(expectedSequence.size(), sequence.size());
- WALBERLA_CHECK( equal(expectedSequence.begin(),expectedSequence.end(), sequence.begin() ) );
-
- expectedSequence.clear();
- sequence.clear();
-
-
- // ------------ Second Run - GPU and SRT -------------------------------------------
-
-
- expectedSequence.push_back("Pre1");
- expectedSequence.push_back("GPU1");
- expectedSequence.push_back("Post1");
- expectedSequence.push_back("Pre2");
- expectedSequence.push_back("GPU2");
- expectedSequence.push_back("Post2");
- expectedSequence.push_back("PostTimestepGPU");
-
- tl->singleStep(gpuSelect + srtSelect);
-
- WALBERLA_CHECK_EQUAL(expectedSequence.size(), sequence.size());
- WALBERLA_CHECK( equal(expectedSequence.begin(),expectedSequence.end(), sequence.begin() ) );
-
- expectedSequence.clear();
- sequence.clear();
-
-
- return 0;
+ timeloop.run(timingPool);
+ for (auto& block : *blocks)
+ {
+ auto field = block.getData< Field_T >(fieldID);
+ if (block.getAABB().min()[0] < 20)
+ {
+ for (auto iter = field->begin(); iter != field->end(); ++iter)
+ WALBERLA_CHECK_EQUAL(*iter, timesteps)
+ }
+ else
+ {
+ for (auto iter = field->begin(); iter != field->end(); ++iter)
+ WALBERLA_CHECK_EQUAL(*iter, timesteps * 2)
+ }
+ }
+
+ return EXIT_SUCCESS;
}
-
-
-
--
GitLab