20585279183

Committed 29 Dec 2025 11:48PM UTC coverage: 39.581% (-0.8%) from 40.359%

Build # 20585279183

Build Type

push

github

Committed by

web-flow

Commit Message

Merge pull request #412 from daisytuner/mean-std-nodes

adds mean/std library nodes

Run Details

14647 of 48066 branches covered (30.47%)

Branch coverage included in aggregate %.

225 of 622 new or added lines in 14 files covered. (36.17%)

41 existing lines in 1 file now uncovered.

12489 of 20493 relevant lines covered (60.94%)

87.19 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

50.91

/src/data_flow/library_nodes/math/tensor/reduce_ops/std_node.cpp

#include "sdfg/data_flow/library_nodes/math/tensor/reduce_ops/std_node.h"
#include "sdfg/analysis/scope_analysis.h"
#include "sdfg/builder/structured_sdfg_builder.h"
#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/pow_node.h"
#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/sqrt_node.h"
#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/sub_node.h"
#include "sdfg/data_flow/library_nodes/math/tensor/reduce_ops/mean_node.h"
#include "sdfg/data_flow/library_nodes/stdlib/malloc.h"
#include "sdfg/structured_control_flow/block.h"
#include "sdfg/types/scalar.h"
#include "sdfg/types/utils.h"

namespace sdfg {
namespace math {
namespace tensor {

StdNode::StdNode(
    size_t element_id,
    const DebugInfo& debug_info,
    const graph::Vertex vertex,
    data_flow::DataFlowGraph& parent,
    const std::vector<symbolic::Expression>& shape,
    const std::vector<int64_t>& axes,
    bool keepdims
)
    : ReduceNode(element_id, debug_info, vertex, parent, LibraryNodeType_Std, shape, axes, keepdims) {}

bool StdNode::expand(builder::StructuredSDFGBuilder& builder, analysis::AnalysisManager& analysis_manager) {
    auto& dataflow = this->get_parent();
    auto& block = static_cast<structured_control_flow::Block&>(*dataflow.get_parent());

    if (dataflow.in_degree(*this) != 1 || dataflow.out_degree(*this) != 1) {
        return false;
    }

    auto& scope_analysis = analysis_manager.get<analysis::ScopeAnalysis>();
    auto& parent = static_cast<structured_control_flow::Sequence&>(*scope_analysis.parent_scope(&block));
    int index = parent.index(block);
    auto& transition = parent.at(index).second;

    auto& in_edge = *dataflow.in_edges(*this).begin();
    auto& out_edge = *dataflow.out_edges(*this).begin();
    auto& in_node = static_cast<data_flow::AccessNode&>(in_edge.src());
    auto& out_node = static_cast<data_flow::AccessNode&>(out_edge.dst());

    // Calculate output shape
    std::vector<symbolic::Expression> output_shape;
    std::vector<int64_t> sorted_axes = axes_;
    std::sort(sorted_axes.begin(), sorted_axes.end());

    for (size_t i = 0; i < shape_.size(); ++i) {
        bool is_axis = false;
        for (auto axis : sorted_axes) {
            if (axis == (int64_t) i) {
                is_axis = true;
                break;
            }
        }

        if (is_axis) {
            if (keepdims_) {
                output_shape.push_back(symbolic::one());
            }
        } else {
            output_shape.push_back(shape_[i]);
        }
    }

    std::string tmp_x2_name = builder.find_new_name("_std_x2");
    builder.add_container(tmp_x2_name, in_edge.base_type());
    std::string tmp_mean_x2_name = builder.find_new_name("_std_mean_x2");
    builder.add_container(tmp_mean_x2_name, out_edge.base_type());
    std::string tmp_mean_x_name = builder.find_new_name("_std_mean_x");
    builder.add_container(tmp_mean_x_name, out_edge.base_type());

    if (in_edge.base_type().type_id() == types::TypeID::Pointer) {
        auto& pointee_type = static_cast<const types::Pointer&>(in_edge.base_type()).pointee_type();
        symbolic::Expression bytes_in = types::get_type_size(pointee_type, false);
        for (auto& dim : this->shape_) {
            bytes_in = symbolic::mul(dim, bytes_in);
        }

        {
            auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());
            auto& tmp_x2_name_access = builder.add_access(alloc_block, tmp_x2_name);
            auto& tmp_x2_name_malloc_node =
                builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_in);
            builder.add_computational_memlet(
                alloc_block,
                tmp_x2_name_malloc_node,
                "_ret",
                tmp_x2_name_access,
                {},
                in_edge.base_type(),
                this->debug_info()
            );
        }
    }

    if (out_edge.base_type().type_id() == types::TypeID::Pointer) {
        auto& pointee_type = static_cast<const types::Pointer&>(out_edge.base_type()).pointee_type();
        symbolic::Expression bytes_out = types::get_type_size(pointee_type, false);
        for (auto& dim : output_shape) {
            bytes_out = symbolic::mul(dim, bytes_out);
        }

        {
            auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());
            auto& tmp_mean_x2_name_access = builder.add_access(alloc_block, tmp_mean_x2_name);
            auto& tmp_mean_x2_name_malloc_node =
                builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_out);
            builder.add_computational_memlet(
                alloc_block,
                tmp_mean_x2_name_malloc_node,
                "_ret",
                tmp_mean_x2_name_access,
                {},
                in_edge.base_type(),
                this->debug_info()
            );
        }

        {
            auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());
            auto& tmp_mean_x_name_access = builder.add_access(alloc_block, tmp_mean_x_name);
            auto& tmp_mean_x_name_malloc_node =
                builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_out);
            builder.add_computational_memlet(
                alloc_block,
                tmp_mean_x_name_malloc_node,
                "_ret",
                tmp_mean_x_name_access,
                {},
                in_edge.base_type(),
                this->debug_info()
            );
        }
    }

    // 1. X^2
    auto& pow_block = builder.add_block_before(parent, block, {}, this->debug_info());
    auto& pow_in_node = builder.add_access(pow_block, in_node.data(), this->debug_info());
    auto& pow_out_node = builder.add_access(pow_block, tmp_x2_name, this->debug_info());

    auto& pow_node_1 = builder.add_library_node<PowNode>(pow_block, this->debug_info(), shape_);
    auto& const_2 =
        builder.add_constant(pow_block, "2", types::Scalar(types::PrimitiveType::Int64), this->debug_info());

    builder
        .add_computational_memlet(pow_block, pow_in_node, pow_node_1, "A", {}, in_edge.base_type(), this->debug_info());
    builder.add_computational_memlet(
        pow_block, const_2, pow_node_1, "B", {}, types::Scalar(types::PrimitiveType::Int64), this->debug_info()
    );
    builder
        .add_computational_memlet(pow_block, pow_node_1, "C", pow_out_node, {}, in_edge.base_type(), this->debug_info());

    // 2. Mean(X^2)
    auto& mean_x2_block = builder.add_block_before(parent, block, {}, this->debug_info());
    auto& mean_x2_in_node = builder.add_access(mean_x2_block, tmp_x2_name, this->debug_info());
    auto& mean_x2_out_node = builder.add_access(mean_x2_block, tmp_mean_x2_name, this->debug_info());

    auto& mean_node_1 = builder.add_library_node<MeanNode>(mean_x2_block, this->debug_info(), shape_, axes_, keepdims_);
    builder.add_computational_memlet(
        mean_x2_block, mean_x2_in_node, mean_node_1, "X", {}, in_edge.base_type(), this->debug_info()
    );
    builder.add_computational_memlet(
        mean_x2_block, mean_node_1, "Y", mean_x2_out_node, {}, out_edge.base_type(), this->debug_info()
    );

    // 3. Mean(X)
    auto& mean_x_block = builder.add_block_before(parent, block, {}, this->debug_info());
    auto& mean_x_in_node = builder.add_access(mean_x_block, in_node.data(), this->debug_info());
    auto& mean_x_out_node = builder.add_access(mean_x_block, tmp_mean_x_name, this->debug_info());

    auto& mean_node_2 = builder.add_library_node<MeanNode>(mean_x_block, this->debug_info(), shape_, axes_, keepdims_);
    builder.add_computational_memlet(
        mean_x_block, mean_x_in_node, mean_node_2, "X", {}, in_edge.base_type(), this->debug_info()
    );
    builder.add_computational_memlet(
        mean_x_block, mean_node_2, "Y", mean_x_out_node, {}, out_edge.base_type(), this->debug_info()
    );

    // 4. Mean(X)^2
    auto& pow_mean_x_block = builder.add_block_before(parent, block, {}, this->debug_info());
    auto& pow_mean_x_in_node = builder.add_access(pow_mean_x_block, tmp_mean_x_name, this->debug_info());
    auto& pow_mean_x_out_node = builder.add_access(pow_mean_x_block, tmp_mean_x_name, this->debug_info());

    auto& pow_node_2 = builder.add_library_node<PowNode>(pow_mean_x_block, this->debug_info(), output_shape);
    auto& const_2_2 =
        builder.add_constant(pow_mean_x_block, "2", types::Scalar(types::PrimitiveType::Int64), this->debug_info());

    builder.add_computational_memlet(
        pow_mean_x_block, pow_mean_x_in_node, pow_node_2, "A", {}, out_edge.base_type(), this->debug_info()
    );
    builder.add_computational_memlet(
        pow_mean_x_block, const_2_2, pow_node_2, "B", {}, types::Scalar(types::PrimitiveType::Int64), this->debug_info()
    );
    builder.add_computational_memlet(
        pow_mean_x_block, pow_node_2, "C", pow_mean_x_out_node, {}, out_edge.base_type(), this->debug_info()
    );

    // 5. Mean(X^2) - Mean(X)^2
    auto& sub_block = builder.add_block_before(parent, block, {}, this->debug_info());
    auto& sub_in1_node = builder.add_access(sub_block, tmp_mean_x2_name, this->debug_info());
    auto& sub_in2_node = builder.add_access(sub_block, tmp_mean_x_name, this->debug_info());
    auto& sub_out_node = builder.add_access(sub_block, out_node.data(), this->debug_info());

    auto& sub_node = builder.add_library_node<SubNode>(sub_block, this->debug_info(), output_shape);
    builder
        .add_computational_memlet(sub_block, sub_in1_node, sub_node, "A", {}, out_edge.base_type(), this->debug_info());
    builder
        .add_computational_memlet(sub_block, sub_in2_node, sub_node, "B", {}, out_edge.base_type(), this->debug_info());
    builder
        .add_computational_memlet(sub_block, sub_node, "C", sub_out_node, {}, out_edge.base_type(), this->debug_info());

    // 6. Sqrt(...)
    auto& sqrt_block = builder.add_block_before(parent, block, transition.assignments(), this->debug_info());
    auto& sqrt_in_node = builder.add_access(sqrt_block, out_node.data(), this->debug_info());
    auto& sqrt_out_node = builder.add_access(sqrt_block, out_node.data(), this->debug_info());

    auto& sqrt_node = builder.add_library_node<SqrtNode>(sqrt_block, this->debug_info(), output_shape);
    builder
        .add_computational_memlet(sqrt_block, sqrt_in_node, sqrt_node, "X", {}, out_edge.base_type(), this->debug_info());
    builder
        .add_computational_memlet(sqrt_block, sqrt_node, "Y", sqrt_out_node, {}, out_edge.base_type(), this->debug_info());

    // Cleanup
    builder.remove_memlet(block, in_edge);
    builder.remove_memlet(block, out_edge);
    builder.remove_node(block, in_node);
    builder.remove_node(block, out_node);
    builder.remove_node(block, *this);

    int last_index = parent.index(block);
    builder.remove_child(parent, last_index);

    return true;
}

bool StdNode::expand_reduction(
    builder::StructuredSDFGBuilder& builder,
    analysis::AnalysisManager& analysis_manager,
    structured_control_flow::Sequence& body,
    const std::string& input_name,
    const std::string& output_name,
    const types::IType& input_type,
    const types::IType& output_type,
    const data_flow::Subset& input_subset,
    const data_flow::Subset& output_subset
) {
    throw std::runtime_error("StdNode::expand_reduction should not be called");
}

std::string StdNode::identity() const { return "0"; }

std::unique_ptr<data_flow::DataFlowNode> StdNode::
    clone(size_t element_id, const graph::Vertex vertex, data_flow::DataFlowGraph& parent) const {
    return std::unique_ptr<
        data_flow::DataFlowNode>(new StdNode(element_id, debug_info_, vertex, parent, shape_, axes_, keepdims_));
}

} // namespace tensor
} // namespace math
} // namespace sdfg

1	#include "sdfg/data_flow/library_nodes/math/tensor/reduce_ops/std_node.h"
2	#include "sdfg/analysis/scope_analysis.h"
3	#include "sdfg/builder/structured_sdfg_builder.h"
4	#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/pow_node.h"
5	#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/sqrt_node.h"
6	#include "sdfg/data_flow/library_nodes/math/tensor/elementwise_ops/sub_node.h"
7	#include "sdfg/data_flow/library_nodes/math/tensor/reduce_ops/mean_node.h"
8	#include "sdfg/data_flow/library_nodes/stdlib/malloc.h"
9	#include "sdfg/structured_control_flow/block.h"
10	#include "sdfg/types/scalar.h"
11	#include "sdfg/types/utils.h"
12
13	namespace sdfg {
14	namespace math {
15	namespace tensor {
16
17	StdNode::StdNode(	2✔
18	size_t element_id,
19	const DebugInfo& debug_info,
20	const graph::Vertex vertex,
21	data_flow::DataFlowGraph& parent,
22	const std::vector<symbolic::Expression>& shape,
23	const std::vector<int64_t>& axes,
24	bool keepdims
25	)
26	: ReduceNode(element_id, debug_info, vertex, parent, LibraryNodeType_Std, shape, axes, keepdims) {}	2✔
27
28	bool StdNode::expand(builder::StructuredSDFGBuilder& builder, analysis::AnalysisManager& analysis_manager) {	2✔
29	auto& dataflow = this->get_parent();	2✔
30	auto& block = static_cast<structured_control_flow::Block&>(*dataflow.get_parent());	2✔
31
32	if (dataflow.in_degree(this) != 1 \|\| dataflow.out_degree(this) != 1) {	2!
NEW 33	return false;	×
34	}
35
36	auto& scope_analysis = analysis_manager.get<analysis::ScopeAnalysis>();	2✔
37	auto& parent = static_cast<structured_control_flow::Sequence&>(*scope_analysis.parent_scope(&block));	2✔
38	int index = parent.index(block);	2✔
39	auto& transition = parent.at(index).second;	2✔
40
41	auto& in_edge = dataflow.in_edges(this).begin();	2✔
42	auto& out_edge = dataflow.out_edges(this).begin();	2✔
43	auto& in_node = static_cast<data_flow::AccessNode&>(in_edge.src());	2✔
44	auto& out_node = static_cast<data_flow::AccessNode&>(out_edge.dst());	2✔
45
46	// Calculate output shape
47	std::vector<symbolic::Expression> output_shape;	2✔
48	std::vector<int64_t> sorted_axes = axes_;	2!
49	std::sort(sorted_axes.begin(), sorted_axes.end());	2!
50
51	for (size_t i = 0; i < shape_.size(); ++i) {	4✔
52	bool is_axis = false;	2✔
53	for (auto axis : sorted_axes) {	2!
54	if (axis == (int64_t) i) {	2!
55	is_axis = true;	2✔
56	break;	2✔
57	}
58	}
59
60	if (is_axis) {	2!
61	if (keepdims_) {	2!
NEW 62	output_shape.push_back(symbolic::one());	×
NEW 63	}	×
64	} else {	2✔
NEW 65	output_shape.push_back(shape_[i]);	×
66	}
67	}	2✔
68
69	std::string tmp_x2_name = builder.find_new_name("_std_x2");	2!
70	builder.add_container(tmp_x2_name, in_edge.base_type());	2!
71	std::string tmp_mean_x2_name = builder.find_new_name("_std_mean_x2");	2!
72	builder.add_container(tmp_mean_x2_name, out_edge.base_type());	2!
73	std::string tmp_mean_x_name = builder.find_new_name("_std_mean_x");	2!
74	builder.add_container(tmp_mean_x_name, out_edge.base_type());	2!
75
76	if (in_edge.base_type().type_id() == types::TypeID::Pointer) {	2!
77	auto& pointee_type = static_cast<const types::Pointer&>(in_edge.base_type()).pointee_type();	2!
78	symbolic::Expression bytes_in = types::get_type_size(pointee_type, false);	2!
79	for (auto& dim : this->shape_) {	4✔
80	bytes_in = symbolic::mul(dim, bytes_in);	2!
81	}
82
83	{
84	auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
85	auto& tmp_x2_name_access = builder.add_access(alloc_block, tmp_x2_name);	2!
86	auto& tmp_x2_name_malloc_node =	2✔
87	builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_in);	2!
88	builder.add_computational_memlet(	4!
89	alloc_block,	2✔
90	tmp_x2_name_malloc_node,	2✔
91	"_ret",	2!
92	tmp_x2_name_access,	2✔
93	{},	2✔
94	in_edge.base_type(),	2!
95	this->debug_info()	2!
96	);
97	}
98	}	2✔
99
100	if (out_edge.base_type().type_id() == types::TypeID::Pointer) {	2!
NEW 101	auto& pointee_type = static_cast<const types::Pointer&>(out_edge.base_type()).pointee_type();	×
NEW 102	symbolic::Expression bytes_out = types::get_type_size(pointee_type, false);	×
NEW 103	for (auto& dim : output_shape) {	×
NEW 104	bytes_out = symbolic::mul(dim, bytes_out);	×
105	}
106
107	{
NEW 108	auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());	×
NEW 109	auto& tmp_mean_x2_name_access = builder.add_access(alloc_block, tmp_mean_x2_name);	×
NEW 110	auto& tmp_mean_x2_name_malloc_node =	×
NEW 111	builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_out);	×
NEW 112	builder.add_computational_memlet(	×
NEW 113	alloc_block,	×
NEW 114	tmp_mean_x2_name_malloc_node,	×
NEW 115	"_ret",	×
NEW 116	tmp_mean_x2_name_access,	×
NEW 117	{},	×
NEW 118	in_edge.base_type(),	×
NEW 119	this->debug_info()	×
120	);
121	}
122
123	{
NEW 124	auto& alloc_block = builder.add_block_before(parent, block, {}, this->debug_info());	×
NEW 125	auto& tmp_mean_x_name_access = builder.add_access(alloc_block, tmp_mean_x_name);	×
NEW 126	auto& tmp_mean_x_name_malloc_node =	×
NEW 127	builder.add_library_node<stdlib::MallocNode>(alloc_block, this->debug_info(), bytes_out);	×
NEW 128	builder.add_computational_memlet(	×
NEW 129	alloc_block,	×
NEW 130	tmp_mean_x_name_malloc_node,	×
NEW 131	"_ret",	×
NEW 132	tmp_mean_x_name_access,	×
NEW 133	{},	×
NEW 134	in_edge.base_type(),	×
NEW 135	this->debug_info()	×
136	);
137	}
NEW 138	}	×
139
140	// 1. X^2
141	auto& pow_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
142	auto& pow_in_node = builder.add_access(pow_block, in_node.data(), this->debug_info());	2!
143	auto& pow_out_node = builder.add_access(pow_block, tmp_x2_name, this->debug_info());	2!
144
145	auto& pow_node_1 = builder.add_library_node<PowNode>(pow_block, this->debug_info(), shape_);	2!
146	auto& const_2 =	2✔
147	builder.add_constant(pow_block, "2", types::Scalar(types::PrimitiveType::Int64), this->debug_info());	2!
148
149	builder	4✔
150	.add_computational_memlet(pow_block, pow_in_node, pow_node_1, "A", {}, in_edge.base_type(), this->debug_info());	2!
151	builder.add_computational_memlet(	4!
152	pow_block, const_2, pow_node_1, "B", {}, types::Scalar(types::PrimitiveType::Int64), this->debug_info()	2!
153	);
154	builder	4✔
155	.add_computational_memlet(pow_block, pow_node_1, "C", pow_out_node, {}, in_edge.base_type(), this->debug_info());	2!
156
157	// 2. Mean(X^2)
158	auto& mean_x2_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
159	auto& mean_x2_in_node = builder.add_access(mean_x2_block, tmp_x2_name, this->debug_info());	2!
160	auto& mean_x2_out_node = builder.add_access(mean_x2_block, tmp_mean_x2_name, this->debug_info());	2!
161
162	auto& mean_node_1 = builder.add_library_node<MeanNode>(mean_x2_block, this->debug_info(), shape_, axes_, keepdims_);	2!
163	builder.add_computational_memlet(	4!
164	mean_x2_block, mean_x2_in_node, mean_node_1, "X", {}, in_edge.base_type(), this->debug_info()	2!
165	);
166	builder.add_computational_memlet(	4!
167	mean_x2_block, mean_node_1, "Y", mean_x2_out_node, {}, out_edge.base_type(), this->debug_info()	2!
168	);
169
170	// 3. Mean(X)
171	auto& mean_x_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
172	auto& mean_x_in_node = builder.add_access(mean_x_block, in_node.data(), this->debug_info());	2!
173	auto& mean_x_out_node = builder.add_access(mean_x_block, tmp_mean_x_name, this->debug_info());	2!
174
175	auto& mean_node_2 = builder.add_library_node<MeanNode>(mean_x_block, this->debug_info(), shape_, axes_, keepdims_);	2!
176	builder.add_computational_memlet(	4!
177	mean_x_block, mean_x_in_node, mean_node_2, "X", {}, in_edge.base_type(), this->debug_info()	2!
178	);
179	builder.add_computational_memlet(	4!
180	mean_x_block, mean_node_2, "Y", mean_x_out_node, {}, out_edge.base_type(), this->debug_info()	2!
181	);
182
183	// 4. Mean(X)^2
184	auto& pow_mean_x_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
185	auto& pow_mean_x_in_node = builder.add_access(pow_mean_x_block, tmp_mean_x_name, this->debug_info());	2!
186	auto& pow_mean_x_out_node = builder.add_access(pow_mean_x_block, tmp_mean_x_name, this->debug_info());	2!
187
188	auto& pow_node_2 = builder.add_library_node<PowNode>(pow_mean_x_block, this->debug_info(), output_shape);	2!
189	auto& const_2_2 =	2✔
190	builder.add_constant(pow_mean_x_block, "2", types::Scalar(types::PrimitiveType::Int64), this->debug_info());	2!
191
192	builder.add_computational_memlet(	4!
193	pow_mean_x_block, pow_mean_x_in_node, pow_node_2, "A", {}, out_edge.base_type(), this->debug_info()	2!
194	);
195	builder.add_computational_memlet(	4!
196	pow_mean_x_block, const_2_2, pow_node_2, "B", {}, types::Scalar(types::PrimitiveType::Int64), this->debug_info()	2!
197	);
198	builder.add_computational_memlet(	4!
199	pow_mean_x_block, pow_node_2, "C", pow_mean_x_out_node, {}, out_edge.base_type(), this->debug_info()	2!
200	);
201
202	// 5. Mean(X^2) - Mean(X)^2
203	auto& sub_block = builder.add_block_before(parent, block, {}, this->debug_info());	2!
204	auto& sub_in1_node = builder.add_access(sub_block, tmp_mean_x2_name, this->debug_info());	2!
205	auto& sub_in2_node = builder.add_access(sub_block, tmp_mean_x_name, this->debug_info());	2!
206	auto& sub_out_node = builder.add_access(sub_block, out_node.data(), this->debug_info());	2!
207
208	auto& sub_node = builder.add_library_node<SubNode>(sub_block, this->debug_info(), output_shape);	2!
209	builder	4✔
210	.add_computational_memlet(sub_block, sub_in1_node, sub_node, "A", {}, out_edge.base_type(), this->debug_info());	2!
211	builder	4✔
212	.add_computational_memlet(sub_block, sub_in2_node, sub_node, "B", {}, out_edge.base_type(), this->debug_info());	2!
213	builder	4✔
214	.add_computational_memlet(sub_block, sub_node, "C", sub_out_node, {}, out_edge.base_type(), this->debug_info());	2!
215
216	// 6. Sqrt(...)
217	auto& sqrt_block = builder.add_block_before(parent, block, transition.assignments(), this->debug_info());	2!
218	auto& sqrt_in_node = builder.add_access(sqrt_block, out_node.data(), this->debug_info());	2!
219	auto& sqrt_out_node = builder.add_access(sqrt_block, out_node.data(), this->debug_info());	2!
220
221	auto& sqrt_node = builder.add_library_node<SqrtNode>(sqrt_block, this->debug_info(), output_shape);	2!
222	builder	4✔
223	.add_computational_memlet(sqrt_block, sqrt_in_node, sqrt_node, "X", {}, out_edge.base_type(), this->debug_info());	2!
224	builder	4✔
225	.add_computational_memlet(sqrt_block, sqrt_node, "Y", sqrt_out_node, {}, out_edge.base_type(), this->debug_info());	2!
226
227	// Cleanup
228	builder.remove_memlet(block, in_edge);	2!
229	builder.remove_memlet(block, out_edge);	2!
230	builder.remove_node(block, in_node);	2!
231	builder.remove_node(block, out_node);	2!
232	builder.remove_node(block, *this);	2!
233
234	int last_index = parent.index(block);	2!
235	builder.remove_child(parent, last_index);	2!
236
237	return true;	2✔
238	}	2✔
239
NEW 240	bool StdNode::expand_reduction(	×
241	builder::StructuredSDFGBuilder& builder,
242	analysis::AnalysisManager& analysis_manager,
243	structured_control_flow::Sequence& body,
244	const std::string& input_name,
245	const std::string& output_name,
246	const types::IType& input_type,
247	const types::IType& output_type,
248	const data_flow::Subset& input_subset,
249	const data_flow::Subset& output_subset
250	) {
NEW 251	throw std::runtime_error("StdNode::expand_reduction should not be called");	×
NEW 252	}	×
253
NEW 254	std::string StdNode::identity() const { return "0"; }	×
255
256	std::unique_ptr<data_flow::DataFlowNode> StdNode::
NEW 257	clone(size_t element_id, const graph::Vertex vertex, data_flow::DataFlowGraph& parent) const {	×
NEW 258	return std::unique_ptr<	×
NEW 259	data_flow::DataFlowNode>(new StdNode(element_id, debug_info_, vertex, parent, shape_, axes_, keepdims_));	×
NEW 260	}	×
261
262	} // namespace tensor
263	} // namespace math
264	} // namespace sdfg

daisytuner / sdfglib / 20585279183

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous