diff --git a/apps/benchmarks/TurbulentChannel/TurbulentChannel.cpp b/apps/benchmarks/TurbulentChannel/TurbulentChannel.cpp
index a552ea5864bb3c5ee697bb3c15487e198beca9b7..ca78e9f04dad09bce9b12098d2278a1484f146ce 100644
--- a/apps/benchmarks/TurbulentChannel/TurbulentChannel.cpp
+++ b/apps/benchmarks/TurbulentChannel/TurbulentChannel.cpp
@@ -424,11 +424,11 @@ namespace walberla {
ForceCalculator<VectorField_T> const * const forceCalculator,
const std::weak_ptr<StructuredBlockStorage> & blocks,
const BlockDataID velocityFieldId, const BlockDataID meanVelocityFieldId,
- const BlockDataID meanTkeSGSFieldId, Welford_T * velocityWelford,
+ const BlockDataID meanTkeSGSFieldId, const BlockDataID sosFieldId, Welford_T * velocityWelford,
const bool separateFile = false)
: parameters_(parameters), forceCalculator_(forceCalculator), timeloop_(timeloop), blocks_(blocks),
velocityWelford_(velocityWelford), velocityFieldId_(velocityFieldId), meanVelocityFieldId_(meanVelocityFieldId),
- meanTkeSGSFieldId_(meanTkeSGSFieldId), plotFrequency_(parameters->plotFrequency), plotStart_(parameters->plotStart),
+ meanTkeSGSFieldId_(meanTkeSGSFieldId), sosFieldId_(sosFieldId), plotFrequency_(parameters->plotFrequency), plotStart_(parameters->plotStart),
separateFiles_(separateFile)
{
if(!plotFrequency_)
@@ -526,7 +526,7 @@ namespace walberla {
const auto * const tkeSGS = block->template getData<ScalarField_T>(meanTkeSGSFieldId_);
WALBERLA_CHECK_NOT_NULLPTR(tkeSGS)
- const auto * const sos = block->template getData<TensorField_T>(velocityWelford_->sum_of_squares_fieldID);
+ const auto * const sos = block->template getData<TensorField_T>(sosFieldId_);
WALBERLA_CHECK_NOT_NULLPTR(sos)
for(uint_t idx = 0; idx < parameters_->channelHalfWidth; ++idx) {
@@ -541,7 +541,7 @@ namespace walberla {
meanVelocityVector[idx] = meanVelocity->get(localCell, flowAxis);
tkeSGSVector[idx] = tkeSGS->get(localCell);
for(uint_t i = 0; i < TensorField_T::F_SIZE; ++i) {
- reynoldsStressVector[idx*TensorField_T::F_SIZE+i] = sos->get(localCell,i) / velocityWelford_->counter_;
+ reynoldsStressVector[idx*TensorField_T::F_SIZE+i] = sos->get(localCell,i) / velocityWelford_->getCounter();
}
tkeResolvedVector[idx] = real_c(0.5) * (
reynoldsStressVector[idx*TensorField_T::F_SIZE+0] +
@@ -622,6 +622,7 @@ namespace walberla {
const BlockDataID velocityFieldId_{};
const BlockDataID meanVelocityFieldId_{};
const BlockDataID meanTkeSGSFieldId_{};
+ const BlockDataID sosFieldId_{};
const uint_t plotFrequency_{};
const uint_t plotStart_{};
@@ -836,16 +837,15 @@ namespace walberla {
communication.addPackInfo(make_shared< PackInfo_T >(pdfFieldId));
auto setNewForce = [&](const real_t newForce) {
- streamCollideSweep.F_x_ = newForce;
- tkeSgsWriter.F_x_ = newForce;
- tkeSgsWriter.F_x_ = newForce;
+ streamCollideSweep.setF_x(newForce);
+ tkeSgsWriter.setF_x(newForce);
};
// plotting
const bool outputSeparateFiles = channelParameter.getParameter<bool>("separate_files", false);
const TurbulentChannelPlotter<WelfordSweep_T > plotter(&simulationParameters, &timeloop, &forceCalculator, blocks,
velocityFieldId, meanVelocityFieldId,
- meanTkeSgsFieldId, &welfordSweep,
+ meanTkeSgsFieldId, sosFieldId, &welfordSweep,
outputSeparateFiles);
//NOTE must convert sweeps that are altered to lambdas, otherwise copy and counter will stay 0
@@ -876,11 +876,11 @@ namespace walberla {
timeloop.add() << Sweep(wfbLambda, "wall function bounce");
timeloop.add() << Sweep(streamCollideLambda, "stream and collide");
timeloop.add() << BeforeFunction([&](){
- const uint_t velCtr = uint_c(welfordSweep.counter_);
+ const uint_t velCtr = uint_c(welfordSweep.getCounter());
if((timeloop.getCurrentTimeStep() == simulationParameters.samplingStart) ||
(timeloop.getCurrentTimeStep() > simulationParameters.samplingStart && simulationParameters.samplingInterval && (velCtr % simulationParameters.samplingInterval == 0))) {
- welfordSweep.counter_ = real_t(0);
- welfordTKESweep.counter_ = real_t(0);
+ welfordSweep.setCounter(real_c(0.0));
+ welfordTKESweep.setCounter(real_c(0.0));
for(auto & block : *blocks) {
auto * sopField = block.template getData<TensorField_T >(sosFieldId);
sopField->setWithGhostLayer(0.0);
@@ -890,8 +890,8 @@ namespace walberla {
}
}
- welfordSweep.counter_ = welfordSweep.counter_ + real_c(1);
- welfordTKESweep.counter_ = welfordTKESweep.counter_ + real_c(1);
+ welfordSweep.setCounter(welfordSweep.getCounter() + real_c(1.0));
+ welfordTKESweep.setCounter(welfordTKESweep.getCounter() + real_c(1.0));
}, "welford sweep")
<< Sweep(welfordLambda, "welford sweep");
timeloop.add() << Sweep(tkeSgsWriter, "TKE_SGS writer");
diff --git a/python/lbmpy_walberla/templates/LBMSweepCollection.tmpl.h b/python/lbmpy_walberla/templates/LBMSweepCollection.tmpl.h
index 9471b6f29bdf7951199b34b87e593fb4622e46d2..a85b9a93bbb676824b7e82ce3a075cbde1979925 100644
--- a/python/lbmpy_walberla/templates/LBMSweepCollection.tmpl.h
+++ b/python/lbmpy_walberla/templates/LBMSweepCollection.tmpl.h
@@ -450,6 +450,9 @@ class {{class_name}}
}
{%endif%}
+ {{kernel_list|generate_getter(parameter_registration=parameter_scaling)|indent(3)}}
+ {{kernel_list|generate_setter(parameter_registration=parameter_scaling)|indent(3)}}
+
private:
shared_ptr< StructuredBlockForest > blocks_;
{{kernel_list|generate_members(parameter_registration=parameter_scaling)|indent(4)}}
diff --git a/python/pystencils_walberla/jinja_filters.py b/python/pystencils_walberla/jinja_filters.py
index d83ed6bba4197659ae447f773092111bcf7298ec..22ae0eceaf6cd78b5a4f9663c8c20cbe4a915966 100644
--- a/python/pystencils_walberla/jinja_filters.py
+++ b/python/pystencils_walberla/jinja_filters.py
@@ -106,6 +106,35 @@ def get_field_stride(param):
return strides[param.symbol.coordinate]
+def normalise_cell_interval(cell_interval, field_object):
+ if isinstance(cell_interval, str):
+ return cell_interval
+ elif isinstance(cell_interval, dict):
+ return cell_interval[field_object.name]
+ else:
+ return None
+
+
+def get_start_coordinates(cell_interval, field_object, ghost_layers_to_include, required_ghost_layers):
+ ci = normalise_cell_interval(cell_interval, field_object)
+ if ci is None:
+ return [-ghost_layers_to_include - required_ghost_layers] * field_object.spatial_dimensions
+ else:
+ assert ghost_layers_to_include == 0
+ return [sp.Symbol(f"{ci}.{coord_name}Min()") - required_ghost_layers for coord_name in ('x', 'y', 'z')]
+
+def get_end_coordinates(cell_interval, field_object, ghost_layers_to_include, required_ghost_layers):
+ ci = normalise_cell_interval(cell_interval, field_object)
+ if ci is None:
+ shape_names = ['xSize()', 'ySize()', 'zSize()'][:field_object.spatial_dimensions]
+ offset = 2 * ghost_layers_to_include + 2 * required_ghost_layers
+ return [f"int64_c({field_object.name}->{e}) + {offset}" for e in shape_names]
+ else:
+ assert ghost_layers_to_include == 0
+ return [f"int64_c({ci}.{coord_name}Size()) + {2 * required_ghost_layers}"
+ for coord_name in ('x', 'y', 'z')]
+
+
def generate_declaration(kernel_info, target=Target.CPU):
"""Generates the declaration of the kernel function"""
target = translate_target(target)
@@ -308,34 +337,6 @@ def generate_call(ctx, kernel, ghost_layers_to_include=0, cell_interval=None, st
required_ghost_layers = max(max(kernel_ghost_layers))
kernel_call_lines = []
-
- def get_cell_interval(field_object):
- if isinstance(cell_interval, str):
- return cell_interval
- elif isinstance(cell_interval, dict):
- return cell_interval[field_object.name]
- else:
- return None
-
- def get_start_coordinates(field_object):
- ci = get_cell_interval(field_object)
- if ci is None:
- return [-ghost_layers_to_include - required_ghost_layers] * field_object.spatial_dimensions
- else:
- assert ghost_layers_to_include == 0
- return [sp.Symbol(f"{ci}.{coord_name}Min()") - required_ghost_layers for coord_name in ('x', 'y', 'z')]
-
- def get_end_coordinates(field_object):
- ci = get_cell_interval(field_object)
- if ci is None:
- shape_names = ['xSize()', 'ySize()', 'zSize()'][:field_object.spatial_dimensions]
- offset = 2 * ghost_layers_to_include + 2 * required_ghost_layers
- return [f"int64_c({field_object.name}->{e}) + {offset}" for e in shape_names]
- else:
- assert ghost_layers_to_include == 0
- return [f"int64_c({ci}.{coord_name}Size()) + {2 * required_ghost_layers}"
- for coord_name in ('x', 'y', 'z')]
-
for param in ast_params:
if param.is_field_parameter and FieldType.is_indexed(param.fields[0]):
continue
@@ -345,7 +346,7 @@ def generate_call(ctx, kernel, ghost_layers_to_include=0, cell_interval=None, st
if field.field_type == FieldType.BUFFER:
kernel_call_lines.append(f"{param.symbol.dtype} {param.symbol.name} = {param.field_name};")
else:
- coordinates = get_start_coordinates(field)
+ coordinates = get_start_coordinates(cell_interval, field, ghost_layers_to_include, required_ghost_layers)
actual_gls = f"int_c({param.field_name}->nrOfGhostLayers())"
coord_set = set(coordinates)
coord_set = sorted(coord_set, key=lambda e: str(e))
@@ -373,7 +374,8 @@ def generate_call(ctx, kernel, ghost_layers_to_include=0, cell_interval=None, st
coord = param.symbol.coordinate
field = param.fields[0]
type_str = param.symbol.dtype.c_name
- shape = f"{type_str}({get_end_coordinates(field)[coord]})"
+ end_coord = get_end_coordinates(cell_interval, field, ghost_layers_to_include, required_ghost_layers)[coord]
+ shape = f"{type_str}({end_coord})"
assert coord < 3
max_value = f"{field.name}->{('x', 'y', 'z')[coord]}SizeWithGhostLayer()"
kernel_call_lines.append(f"WALBERLA_ASSERT_GREATER_EQUAL({max_value}, {shape})")
@@ -711,8 +713,47 @@ def nested_class_method_definition_prefix(ctx, nested_class_name):
return f"{outer_class}::{nested_class_name}"
-@jinja2_context_decorator
-def generate_parameter_registration(ctx, kernel_infos, parameter_registration):
+def generate_getter(kernel_infos, parameter_registration=None):
+ if not isinstance(kernel_infos, Iterable):
+ kernel_infos = [kernel_infos]
+
+ params_to_skip = []
+ result = []
+ for kernel_info in kernel_infos:
+ for param in kernel_info.parameters:
+ if not param.is_field_parameter and param.symbol.name not in params_to_skip:
+ if parameter_registration and param.symbol.name in parameter_registration.scaling_info:
+ continue
+ dtype = param.symbol.dtype
+ name = param.symbol.name
+ code_line = f"inline {dtype} get{name.capitalize()}() const {{ return {name}_; }}"
+ result.append(code_line)
+ params_to_skip.append(param.symbol.name)
+
+ return "\n".join(result)
+
+
+def generate_setter(kernel_infos, parameter_registration=None):
+ if not isinstance(kernel_infos, Iterable):
+ kernel_infos = [kernel_infos]
+
+ params_to_skip = []
+ result = []
+ for kernel_info in kernel_infos:
+ for param in kernel_info.parameters:
+ if not param.is_field_parameter and param.symbol.name not in params_to_skip:
+ if parameter_registration and param.symbol.name in parameter_registration.scaling_info:
+ continue
+ dtype = param.symbol.dtype
+ name = param.symbol.name
+ code_line = f"inline void set{name.capitalize()}(const {dtype} value) {{ {name}_ = value; }}"
+ result.append(code_line)
+ params_to_skip.append(param.symbol.name)
+
+ return "\n".join(result)
+
+
+def generate_parameter_registration(kernel_infos, parameter_registration):
if parameter_registration is None:
return ""
if not isinstance(kernel_infos, Iterable):
@@ -731,8 +772,7 @@ def generate_parameter_registration(ctx, kernel_infos, parameter_registration):
return "\n".join(result)
-@jinja2_context_decorator
-def generate_constructor(ctx, kernel_infos, parameter_registration):
+def generate_constructor(kernel_infos, parameter_registration):
if parameter_registration is None:
return ""
if not isinstance(kernel_infos, Iterable):
@@ -752,8 +792,7 @@ def generate_constructor(ctx, kernel_infos, parameter_registration):
return "\n".join(result)
-@jinja2_context_decorator
-def generate_field_strides(ctx, kernel_info):
+def generate_field_strides(kernel_info):
result = []
for param in kernel_info.parameters:
if param.is_field_stride:
@@ -835,3 +874,5 @@ def add_pystencils_filters_to_jinja_env(jinja_env):
jinja_env.filters['list_of_expressions'] = generate_list_of_expressions
jinja_env.filters['field_type'] = field_type
jinja_env.filters['generate_field_strides'] = generate_field_strides
+ jinja_env.filters['generate_getter'] = generate_getter
+ jinja_env.filters['generate_setter'] = generate_setter
diff --git a/python/pystencils_walberla/templates/Sweep.tmpl.h b/python/pystencils_walberla/templates/Sweep.tmpl.h
index f0f802288e4dd3a8b5835c7f5ad4960892ad6145..bebbe855c33cf00fc825ca51d5a057f05fa079a1 100644
--- a/python/pystencils_walberla/templates/Sweep.tmpl.h
+++ b/python/pystencils_walberla/templates/Sweep.tmpl.h
@@ -26,9 +26,7 @@
{%- elif target is equalto 'gpu' -%}
#include "gpu/GPUField.h"
#include "gpu/GPUWrapper.h"
-{% if inner_outer_split -%}
-#include "gpu/ParallelStreams.h"
-{%- endif %}
+{%if inner_outer_split %} #include "gpu/ParallelStreams.h" {% endif %}
{%- endif %}
#include "field/SwapableCompare.h"
#include "domain_decomposition/BlockDataID.h"
@@ -132,7 +130,7 @@ public:
configured_ = true;
}
{% else %}
- void configure( const shared_ptr<StructuredBlockStorage> & blocks, IBlock * block ){}
+ void configure( const shared_ptr<StructuredBlockStorage> & /*blocks*/, IBlock * /*block*/ ){}
{% endif %}
{% if inner_outer_split %}
@@ -161,23 +159,23 @@ public:
{%endif%}
}
- {{kernel|generate_members|indent(3)}}
+ {% endif %}
-private:
+ {{kernel|generate_getter()|indent(3)}}
+ {{kernel|generate_setter()|indent(3)}}
- {%if target is equalto 'gpu' %} gpu::ParallelStreams parallelStreams_; {% endif %}
+private:
+ {%if target is equalto 'gpu' and inner_outer_split %} gpu::ParallelStreams parallelStreams_; {% endif %}
+ {{kernel|generate_members|indent(3)}}
+ {% if inner_outer_split %}
Cell outerWidth_;
std::vector<CellInterval> layers_;
- {% if block_offset -%}
- bool configured_;
{% endif %}
- {% else %}
- {{ kernel|generate_members|indent(3) }}
+
{% if block_offset -%}
bool configured_;
{% endif %}
-{% endif %}
};
diff --git a/python/pystencils_walberla/templates/SweepCollection.tmpl.h b/python/pystencils_walberla/templates/SweepCollection.tmpl.h
index 2a89fef5e4c6c825443db51ed1acc3a6d02ffaa8..004f975b4ed98bf1a0b8026821ed72c2792cc3fb 100644
--- a/python/pystencils_walberla/templates/SweepCollection.tmpl.h
+++ b/python/pystencils_walberla/templates/SweepCollection.tmpl.h
@@ -287,6 +287,9 @@ class {{class_name}}
}
{%endif%}
+ {{kernel_list|generate_getter(parameter_registration=parameter_scaling)|indent(3)}}
+ {{kernel_list|generate_setter(parameter_registration=parameter_scaling)|indent(3)}}
+
private:
shared_ptr< StructuredBlockStorage > blocks_;
{{kernel_list|generate_members(parameter_registration=parameter_scaling)|indent(4)}}