15852980623

Committed 24 Jun 2025 02:16PM UTC coverage: 64.412% (+0.3%) from 64.145%

Build # 15852980623

Build Type

push

github

Committed by

web-flow

Commit Message

Merge pull request #72 from daisytuner/capture-instrumentation

Capture instrumentation

Run Details

363 of 446 new or added lines in 19 files covered. (81.39%)

100 existing lines in 5 files now uncovered.

8389 of 13024 relevant lines covered (64.41%)

116.79 hits per line

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

61.18

/src/codegen/code_generators/cuda_code_generator.cpp

#include "sdfg/codegen/code_generators/cuda_code_generator.h"

#include "sdfg/codegen/dispatchers/node_dispatcher_registry.h"
#include "sdfg/codegen/instrumentation/instrumentation.h"
#include "sdfg/codegen/instrumentation/outermost_loops_instrumentation.h"

namespace sdfg {
namespace codegen {

CUDACodeGenerator::CUDACodeGenerator(StructuredSDFG& sdfg,
                                     InstrumentationStrategy instrumentation_strategy,
                                     bool capture_args_results)
    : CodeGenerator(sdfg, instrumentation_strategy, capture_args_results) {
    if (sdfg.type() != FunctionType_NV_GLOBAL) {
        throw std::runtime_error("CUDACodeGenerator can only be used for GPU SDFGs");
    }
    if (capture_args_results) {
        std::cerr << "CUDACodeGenerator does not support capturing args/results!";
    }
};

bool CUDACodeGenerator::generate() {
    this->dispatch_includes();
    this->dispatch_structures();
    this->dispatch_globals();
    this->dispatch_schedule();
    return true;
};

std::string CUDACodeGenerator::function_definition() {
    /********** Arglist **********/
    std::vector<std::string> args;
    for (auto& container : sdfg_.arguments()) {
        args.push_back(language_extension_.declaration(container, sdfg_.type(container)));
    }
    std::stringstream arglist;
    arglist << sdfg::helpers::join(args, ", ");

    return "extern \"C\" __global__ void " + sdfg_.name() + "(" + arglist.str() + ")";
};

bool CUDACodeGenerator::as_source(const std::filesystem::path& header_path,
                                  const std::filesystem::path& source_path,
                                  const std::filesystem::path& library_path) {
    std::ofstream ofs_header(header_path, std::ofstream::out);
    if (!ofs_header.is_open()) {
        return false;
    }

    std::ofstream ofs_source(source_path, std::ofstream::out);
    if (!ofs_source.is_open()) {
        return false;
    }

    std::ofstream ofs_library(library_path, std::ofstream::out);
    if (!ofs_library.is_open()) {
        return false;
    }

    ofs_header << "#pragma once" << std::endl;
    ofs_header << this->includes_stream_.str() << std::endl;
    ofs_header << this->classes_stream_.str() << std::endl;
    ofs_header.close();

    ofs_source << "#include \"" << header_path.filename().string() << "\"" << std::endl;
    ofs_source << this->globals_stream_.str() << std::endl;
    ofs_source << this->function_definition() << std::endl;
    ofs_source << "{" << std::endl;
    ofs_source << this->main_stream_.str() << std::endl;
    ofs_source << "}" << std::endl;
    ofs_source.close();

    auto library_content = this->library_stream_.str();
    if (library_content.empty()) {
        ofs_library.close();
        return true;
    }

    ofs_library << "#include \"" << header_path.filename().string() << "\"" << std::endl;
    ofs_library << library_content << std::endl;
    ofs_library.close();

    return true;
};

void CUDACodeGenerator::dispatch_includes() {
    this->includes_stream_ << "#define "
                           << "__DAISY_NVVM__" << std::endl;
    this->includes_stream_ << "#include "
                           << "\"daisyrtl.h\"" << std::endl;
    if (instrumentation_strategy_ != InstrumentationStrategy::NONE)
        this->includes_stream_ << "#include <daisy_rtl.h>" << std::endl;

    this->includes_stream_ << "#define __daisy_min(a,b) ((a)<(b)?(a):(b))" << std::endl;
    this->includes_stream_ << "#define __daisy_max(a,b) ((a)>(b)?(a):(b))" << std::endl;
    this->includes_stream_ << "#define __daisy_fma(a,b,c) a * b + c" << std::endl;
};

void CUDACodeGenerator::dispatch_structures() {
    // Forward declarations
    for (auto& structure : sdfg_.structures()) {
        this->classes_stream_ << "struct " << structure << ";" << std::endl;
    }

    // Generate topology-sorted structure definitions
    typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::bidirectionalS> structures_graph;
    typedef boost::graph_traits<structures_graph>::vertex_descriptor Vertex;
    std::vector<std::string> names;
    for (auto& structure : sdfg_.structures()) {
        names.push_back(structure);
    }
    structures_graph graph(names.size());

    for (auto& structure : names) {
        auto& definition = sdfg_.structure(structure);
        for (size_t i = 0; i < definition.num_members(); i++) {
            auto member_type = &definition.member_type(symbolic::integer(i));
            while (dynamic_cast<const types::Array*>(member_type)) {
                auto array_type = static_cast<const types::Array*>(member_type);
                member_type = &array_type->element_type();
            }

            if (auto member_structure = dynamic_cast<const sdfg::types::Structure*>(member_type)) {
                boost::add_edge(
                    std::find(names.begin(), names.end(), member_structure->name()) - names.begin(),
                    std::find(names.begin(), names.end(), structure) - names.begin(), graph);
            }
        }
    }

    std::list<Vertex> order;
    std::unordered_map<Vertex, boost::default_color_type> vertex_colors;
    boost::topological_sort(graph, std::back_inserter(order),
                            boost::color_map(boost::make_assoc_property_map(vertex_colors)));
    order.reverse();

    for (auto& structure_index : order) {
        std::string structure = names.at(structure_index);
        auto& definition = sdfg_.structure(structure);
        this->classes_stream_ << "struct ";
        if (definition.is_packed()) {
            this->classes_stream_ << "__attribute__((packed)) ";
        }
        this->classes_stream_ << structure << std::endl;
        this->classes_stream_ << "{\n";

        for (size_t i = 0; i < definition.num_members(); i++) {
            auto& member_type = definition.member_type(symbolic::integer(i));
            if (dynamic_cast<const sdfg::types::Structure*>(&member_type)) {
                this->classes_stream_ << "struct ";
            }
            this->classes_stream_ << language_extension_.declaration("member_" + std::to_string(i),
                                                                     member_type, false, true);
            this->classes_stream_ << ";" << std::endl;
        }

        this->classes_stream_ << "};" << std::endl;
    }
};

void CUDACodeGenerator::dispatch_globals() {
    for (auto& container : sdfg_.externals()) {
        auto& type = sdfg_.type(container);
        if (type.storage_type() == types::StorageType_NV_Global) {
            this->globals_stream_ << "extern " << language_extension_.declaration(container, type)
                                  << ";" << std::endl;
        }
        if (type.storage_type() == types::StorageType_NV_Constant) {
            assert(type.initializer().empty());
            this->globals_stream_ << "__constant__ "
                                  << language_extension_.declaration(container, type, true) << ";"
                                  << std::endl;
        }
    }
};

void CUDACodeGenerator::dispatch_schedule() {
    // Declare shared memory
    for (auto& container : sdfg_.externals()) {
        auto& type = sdfg_.type(container);
        if (type.storage_type() == types::StorageType_NV_Shared) {
            this->main_stream_ << language_extension_.declaration(container, sdfg_.type(container))
                               << ";" << std::endl;
        }
    }

    // Map external variables to internal variables
    for (auto& container : sdfg_.containers()) {
        if (!sdfg_.is_internal(container)) {
            continue;
        }

        std::string external_name =
            container.substr(0, container.length() - external_suffix.length());
        this->main_stream_ << language_extension_.declaration(container, sdfg_.type(container));
        this->main_stream_ << " = "
                           << "&" << external_name;
        this->main_stream_ << ";" << std::endl;
    }

    // Declare transient containers
    for (auto& container : sdfg_.containers()) {
        if (!sdfg_.is_transient(container)) {
            continue;
        }

        std::string val =
            this->language_extension_.declaration(container, sdfg_.type(container), false, true);
        if (!val.empty()) {
            this->main_stream_ << val;
            this->main_stream_ << ";" << std::endl;
        }
    }

    // Add instrumentation
    auto instrumentation = create_instrumentation(instrumentation_strategy_, sdfg_);

    auto dispatcher = create_dispatcher(language_extension_, sdfg_, sdfg_.root(), *instrumentation);
    dispatcher->dispatch(this->main_stream_, this->globals_stream_, this->library_stream_);
};

}  // namespace codegen
}  // namespace sdfg

1	#include "sdfg/codegen/code_generators/cuda_code_generator.h"
2
3	#include "sdfg/codegen/dispatchers/node_dispatcher_registry.h"
4	#include "sdfg/codegen/instrumentation/instrumentation.h"
5	#include "sdfg/codegen/instrumentation/outermost_loops_instrumentation.h"
6
7	namespace sdfg {
8	namespace codegen {
9
10	CUDACodeGenerator::CUDACodeGenerator(StructuredSDFG& sdfg,	10✔
11	InstrumentationStrategy instrumentation_strategy,
12	bool capture_args_results)
13	: CodeGenerator(sdfg, instrumentation_strategy, capture_args_results) {	5✔
14	if (sdfg.type() != FunctionType_NV_GLOBAL) {	5✔
15	throw std::runtime_error("CUDACodeGenerator can only be used for GPU SDFGs");	×
16	}
17	if (capture_args_results) {	5✔
NEW 18	std::cerr << "CUDACodeGenerator does not support capturing args/results!";	×
NEW 19	}	×
20	};	5✔
21
22	bool CUDACodeGenerator::generate() {	4✔
23	this->dispatch_includes();	4✔
24	this->dispatch_structures();	4✔
25	this->dispatch_globals();	4✔
26	this->dispatch_schedule();	4✔
27	return true;	4✔
28	};
29
30	std::string CUDACodeGenerator::function_definition() {	1✔
31	/******** Arglist ********/
32	std::vector<std::string> args;	1✔
33	for (auto& container : sdfg_.arguments()) {	1✔
34	args.push_back(language_extension_.declaration(container, sdfg_.type(container)));	×
35	}
36	std::stringstream arglist;	1✔
37	arglist << sdfg::helpers::join(args, ", ");	1✔
38
39	return "extern \"C\" __global__ void " + sdfg_.name() + "(" + arglist.str() + ")";	1✔
40	};	1✔
41
42	bool CUDACodeGenerator::as_source(const std::filesystem::path& header_path,	×
43	const std::filesystem::path& source_path,
44	const std::filesystem::path& library_path) {
45	std::ofstream ofs_header(header_path, std::ofstream::out);	×
46	if (!ofs_header.is_open()) {	×
47	return false;	×
48	}
49
50	std::ofstream ofs_source(source_path, std::ofstream::out);	×
51	if (!ofs_source.is_open()) {	×
52	return false;	×
53	}
54
55	std::ofstream ofs_library(library_path, std::ofstream::out);	×
56	if (!ofs_library.is_open()) {	×
57	return false;	×
58	}
59
60	ofs_header << "#pragma once" << std::endl;	×
61	ofs_header << this->includes_stream_.str() << std::endl;	×
62	ofs_header << this->classes_stream_.str() << std::endl;	×
63	ofs_header.close();	×
64
65	ofs_source << "#include \"" << header_path.filename().string() << "\"" << std::endl;	×
66	ofs_source << this->globals_stream_.str() << std::endl;	×
67	ofs_source << this->function_definition() << std::endl;	×
68	ofs_source << "{" << std::endl;	×
69	ofs_source << this->main_stream_.str() << std::endl;	×
70	ofs_source << "}" << std::endl;	×
71	ofs_source.close();	×
72
73	auto library_content = this->library_stream_.str();	×
74	if (library_content.empty()) {	×
75	ofs_library.close();	×
76	return true;	×
77	}
78
79	ofs_library << "#include \"" << header_path.filename().string() << "\"" << std::endl;	×
80	ofs_library << library_content << std::endl;	×
81	ofs_library.close();	×
82
83	return true;	×
84	};	×
85
86	void CUDACodeGenerator::dispatch_includes() {	4✔
87	this->includes_stream_ << "#define "	4✔
88	<< "__DAISY_NVVM__" << std::endl;	4✔
89	this->includes_stream_ << "#include "	4✔
90	<< "\"daisyrtl.h\"" << std::endl;	4✔
91	if (instrumentation_strategy_ != InstrumentationStrategy::NONE)	4✔
92	this->includes_stream_ << "#include <daisy_rtl.h>" << std::endl;	×
93
94	this->includes_stream_ << "#define __daisy_min(a,b) ((a)<(b)?(a):(b))" << std::endl;	4✔
95	this->includes_stream_ << "#define __daisy_max(a,b) ((a)>(b)?(a):(b))" << std::endl;	4✔
96	this->includes_stream_ << "#define __daisy_fma(a,b,c) a * b + c" << std::endl;	4✔
97	};	4✔
98
99	void CUDACodeGenerator::dispatch_structures() {	4✔
100	// Forward declarations
101	for (auto& structure : sdfg_.structures()) {	7✔
102	this->classes_stream_ << "struct " << structure << ";" << std::endl;	3✔
103	}
104
105	// Generate topology-sorted structure definitions
106	typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::bidirectionalS> structures_graph;
107	typedef boost::graph_traits<structures_graph>::vertex_descriptor Vertex;
108	std::vector<std::string> names;	4✔
109	for (auto& structure : sdfg_.structures()) {	7✔
110	names.push_back(structure);	3✔
111	}
112	structures_graph graph(names.size());	4✔
113
114	for (auto& structure : names) {	7✔
115	auto& definition = sdfg_.structure(structure);	3✔
116	for (size_t i = 0; i < definition.num_members(); i++) {	6✔
117	auto member_type = &definition.member_type(symbolic::integer(i));	3✔
118	while (dynamic_cast<const types::Array*>(member_type)) {	3✔
119	auto array_type = static_cast<const types::Array*>(member_type);	×
120	member_type = &array_type->element_type();	×
121	}
122
123	if (auto member_structure = dynamic_cast<const sdfg::types::Structure*>(member_type)) {	3✔
124	boost::add_edge(	1✔
125	std::find(names.begin(), names.end(), member_structure->name()) - names.begin(),	1✔
126	std::find(names.begin(), names.end(), structure) - names.begin(), graph);	1✔
127	}	1✔
128	}	3✔
129	}
130
131	std::list<Vertex> order;	4✔
132	std::unordered_map<Vertex, boost::default_color_type> vertex_colors;	4✔
133	boost::topological_sort(graph, std::back_inserter(order),	8✔
134	boost::color_map(boost::make_assoc_property_map(vertex_colors)));	4✔
135	order.reverse();	4✔
136
137	for (auto& structure_index : order) {	7✔
138	std::string structure = names.at(structure_index);	3✔
139	auto& definition = sdfg_.structure(structure);	3✔
140	this->classes_stream_ << "struct ";	3✔
141	if (definition.is_packed()) {	3✔
142	this->classes_stream_ << "__attribute__((packed)) ";	×
143	}	×
144	this->classes_stream_ << structure << std::endl;	3✔
145	this->classes_stream_ << "{\n";	3✔
146
147	for (size_t i = 0; i < definition.num_members(); i++) {	6✔
148	auto& member_type = definition.member_type(symbolic::integer(i));	3✔
149	if (dynamic_cast<const sdfg::types::Structure*>(&member_type)) {	3✔
150	this->classes_stream_ << "struct ";	1✔
151	}	1✔
152	this->classes_stream_ << language_extension_.declaration("member_" + std::to_string(i),	3✔
153	member_type, false, true);	3✔
154	this->classes_stream_ << ";" << std::endl;	3✔
155	}	3✔
156
157	this->classes_stream_ << "};" << std::endl;	3✔
158	}	3✔
159	};	4✔
160
161	void CUDACodeGenerator::dispatch_globals() {	4✔
162	for (auto& container : sdfg_.externals()) {	5✔
163	auto& type = sdfg_.type(container);	1✔
164	if (type.storage_type() == types::StorageType_NV_Global) {	1✔
165	this->globals_stream_ << "extern " << language_extension_.declaration(container, type)	2✔
166	<< ";" << std::endl;	1✔
167	}	1✔
168	if (type.storage_type() == types::StorageType_NV_Constant) {	1✔
169	assert(type.initializer().empty());	×
170	this->globals_stream_ << "__constant__ "	×
171	<< language_extension_.declaration(container, type, true) << ";"	×
172	<< std::endl;	×
173	}	×
174	}
175	};	4✔
176
177	void CUDACodeGenerator::dispatch_schedule() {	4✔
178	// Declare shared memory
179	for (auto& container : sdfg_.externals()) {	5✔
180	auto& type = sdfg_.type(container);	1✔
181	if (type.storage_type() == types::StorageType_NV_Shared) {	1✔
182	this->main_stream_ << language_extension_.declaration(container, sdfg_.type(container))	×
183	<< ";" << std::endl;	×
184	}	×
185	}
186
187	// Map external variables to internal variables
188	for (auto& container : sdfg_.containers()) {	5✔
189	if (!sdfg_.is_internal(container)) {	1✔
190	continue;	1✔
191	}
192
193	std::string external_name =
194	container.substr(0, container.length() - external_suffix.length());	×
195	this->main_stream_ << language_extension_.declaration(container, sdfg_.type(container));	×
196	this->main_stream_ << " = "	×
197	<< "&" << external_name;	×
198	this->main_stream_ << ";" << std::endl;	×
199	}	×
200
201	// Declare transient containers
202	for (auto& container : sdfg_.containers()) {	5✔
203	if (!sdfg_.is_transient(container)) {	1✔
204	continue;	1✔
205	}
206
207	std::string val =
208	this->language_extension_.declaration(container, sdfg_.type(container), false, true);	×
209	if (!val.empty()) {	×
210	this->main_stream_ << val;	×
211	this->main_stream_ << ";" << std::endl;	×
212	}	×
213	}	×
214
215	// Add instrumentation
216	auto instrumentation = create_instrumentation(instrumentation_strategy_, sdfg_);	4✔
217
218	auto dispatcher = create_dispatcher(language_extension_, sdfg_, sdfg_.root(), *instrumentation);	4✔
219	dispatcher->dispatch(this->main_stream_, this->globals_stream_, this->library_stream_);	4✔
220	};	4✔
221
222	} // namespace codegen
223	} // namespace sdfg

daisytuner / sdfglib / 15852980623

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