Skip to content

Commit

Permalink
Integrated wf2
Browse files Browse the repository at this point in the history
  • Loading branch information
@KavyaVarma
KavyaVarma committed Apr 9, 2024
1 parent bc46995 commit 9d88802
Show file tree
Hide file tree
Showing 21 changed files with 3,272 additions and 0 deletions.
93 changes: 93 additions & 0 deletions workflows/matching/CMakeLists.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,93 @@
# SPDX-License-Identifier: MIT
# Copyright (c) 2023. University of Texas at Austin. All rights reserved.

### Copyright (c) 2023 Advanced Micro Devices, Inc. All rights reserved. ###

cmake_minimum_required(VERSION 3.21 FATAL_ERROR)

project(
pando-wf2
VERSION 1.0.0
DESCRIPTION "Exact Matching Galois implementation for PANDO"
LANGUAGES CXX
)

if(PROJECT_IS_TOP_LEVEL)
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
set(LINKER_SCRIPT ${CMAKE_BINARY_DIR}/linker.ld)
set(CMAKE_CXX_FLAGS_COVERAGE "-g -O0 --coverage -fprofile-info-section -DENABLE_GCOV=1")
set(CMAKE_EXE_LINKER_FLAGS_COVERAGE "--coverage -T ${LINKER_SCRIPT}")
set(CMAKE_SHARED_LINKER_FLAGS_COVERAGE "--coverage -T ${LINKER_SCRIPT}")
set(CMAKE_MODULE_LINKER_FLAGS_COVERAGE "--coverage -T ${LINKER_SCRIPT}")
endif()

# helper functions
include(${pando-lib-galois_SOURCE_DIR}/cmake/PANDOCompilerOptions.cmake)

# disable in-source builds to avoid source tree corruption
if ("${CMAKE_SOURCE_DIR}" STREQUAL "${CMAKE_BINARY_DIR}")
message(FATAL_ERROR
"In-source builds are not supported. Please create a directory for build files and delete CMakeCache.txt and \
the CMakefiles directory."
)
endif()

# default build type
set(DEFAULT_BUILD_TYPE "Release")
if (NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
message(STATUS "Setting build type to default '${DEFAULT_BUILD_TYPE}' as none was specified.")
set(CMAKE_BUILD_TYPE "${DEFAULT_BUILD_TYPE}" CACHE STRING "Choose the type of build." FORCE)
# possible values of build type for cmake-gui
set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo")
endif ()

# target
pando_add_library(pando-wf2-galois)

# headers
include(GenerateExportHeader)
generate_export_header(pando-wf2-galois EXPORT_FILE_NAME ${CMAKE_CURRENT_BINARY_DIR}/include/pando-wf2-galois/export.h)

target_include_directories(pando-wf2-galois
PUBLIC
"$<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>"
"$<BUILD_INTERFACE:${CMAKE_CURRENT_BINARY_DIR}/include/pando-wf2-galois>"
"$<INSTALL_INTERFACE:include>"
PRIVATE
"${CMAKE_CURRENT_SOURCE_DIR}/src"
)

# library properties
set_target_properties(pando-wf2-galois
PROPERTIES
CXX_STANDARD 17
CXX_STANDARD_REQUIRED ON
CXX_EXTENSIONS OFF
CXX_VISIBILITY_PRESET hidden
VISIBILITY_INLINES_HIDDEN ON
POSITION_INDEPENDENT_CODE ON
VERSION "${PROJECT_VERSION}"
SOVERSION "${PROJECT_VERSION_MAJOR}"
EXPORT_NAME pando-wf2-galois
OUTPUT_NAME pando-wf2-galois
PUBLIC_HEADER "${PUBLIC_HEADERS}"
# enable RPATH during both build and installation; use relative RPATH as well
SKIP_BUILD_RPATH FALSE
BUILD_WITH_INSTALL_RPATH FALSE
INSTALL_RPATH_USE_LINK_PATH TRUE
BUILD_RPATH "\$ORIGIN"
INSTALL_RPATH "\$ORIGIN"
# organize lib / bin in the build directory
ARCHIVE_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/lib
LIBRARY_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/lib
RUNTIME_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/bin
)

# sources
add_subdirectory(src)

# tests
option(BUILD_TESTING "Build tests." ON)
if (BUILD_TESTING AND ENABLE_PANDORT_TESTS)
add_subdirectory(test)
endif ()
30 changes: 30 additions & 0 deletions workflows/matching/include/pando-wf2-galois/CPPLINT.cfg
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
# SPDX-License-Identifier: MIT
# Copyright (c) 2023. University of Texas at Austin. All rights reserved.

# Stop searching for additional config files.
set noparent

# Use 'PANDO_WF2_GALOIS_' as the cpp header guard prefix
root=..

# Limit line length.
linelength=100

# We ignore the below rules

# do not want to indent public/private/protected by +1 space inside class
filter=-whitespace/indent
# ok without two spaces between code and comments
filter=-whitespace/comments
filter=-legal/copyright
# do not include the directory when naming header files
filter=-build/include_subdir
filter=-build/c++11
#filter=-build/include_order
# We want to pass non-const references as function arguments
filter=-runtime/references
# cpplint throws false positives for some constexpr expressions that are used for array dims
filter=-runtime/arrays

# Ignore code that may be submodules in future.
# exclude_files=third_party or external
160 changes: 160 additions & 0 deletions workflows/matching/include/pando-wf2-galois/approx_match.h
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,160 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2023. University of Texas at Austin. All rights reserved.
#ifndef PANDO_WF2_GALOIS_APPROX_MATCH_H_
#define PANDO_WF2_GALOIS_APPROX_MATCH_H_

#include "export.h"

#include "pando-rt/export.h"
#include "pando-rt/pando-rt.hpp"
#include "pando-wf2-galois/graph_ds.h"
#include <pando-lib-galois/graphs/dist_local_csr.hpp>

namespace wf2 {
namespace approx {

using Vertex = wf::WMDVertex;
using Edge = wf::WMDEdge;
using Graph = galois::DistLocalCSR<Vertex, Edge>;

struct node_sim {
std::uint64_t lid;
double similarity;
std::uint64_t token;
node_sim(std::uint64_t lid_, double similarity_, std::uint64_t token_)
: lid(lid_), similarity(similarity_), token(token_) {}
node_sim() = default;
};

struct GraphState {
Graph graph;
pando::Vector<std::uint64_t> person_sale_person_bomb_bath;
pando::Vector<std::uint64_t> person_sale_person_pressure_cooker;
pando::Vector<std::uint64_t> person_sale_person_ammunition;
pando::Vector<std::uint64_t> person_sale_person_electronics;
pando::Vector<std::uint64_t> person_purchase_person_bomb_bath;
pando::Vector<std::uint64_t> person_purchase_person_pressure_cooker;
pando::Vector<std::uint64_t> person_purchase_person_ammunition;
pando::Vector<std::uint64_t> person_purchase_person_electronics;
pando::Vector<std::uint64_t> person_author_forumevent;
pando::Vector<std::uint64_t> person_author_publication;
pando::Vector<std::uint64_t> forum_includes_forumevent;
pando::Vector<std::uint64_t> forum_hastopic_topic_nyc;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_bomb;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_explosion;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_williamsburg;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_outdoors;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_prospect_park;
pando::Vector<std::uint64_t> forumevent_hastopic_topic_jihad;
pando::Vector<std::uint64_t> publication_hasorg_topic_near_nyc;
pando::Vector<std::uint64_t> publication_hastopic_topic_electrical_eng;

pando::Vector<pando::Vector<node_sim>> similarity;
pando::Vector<bool> new_matched;
pando::Vector<bool> matched;
pando::Vector<node_sim> match;

explicit GraphState(Graph g_) : graph(g_) {}
GraphState() = default;

void initialize(std::uint64_t other_size) {
std::uint64_t size = graph.size();
PANDO_CHECK(person_sale_person_bomb_bath.initialize(size));
PANDO_CHECK(person_sale_person_pressure_cooker.initialize(size));
PANDO_CHECK(person_sale_person_ammunition.initialize(size));
PANDO_CHECK(person_sale_person_electronics.initialize(size));
PANDO_CHECK(person_purchase_person_bomb_bath.initialize(size));
PANDO_CHECK(person_purchase_person_pressure_cooker.initialize(size));
PANDO_CHECK(person_purchase_person_ammunition.initialize(size));
PANDO_CHECK(person_purchase_person_electronics.initialize(size));
PANDO_CHECK(person_author_forumevent.initialize(size));
PANDO_CHECK(person_author_publication.initialize(size));
PANDO_CHECK(forum_includes_forumevent.initialize(size));
PANDO_CHECK(forum_hastopic_topic_nyc.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_bomb.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_explosion.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_williamsburg.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_outdoors.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_prospect_park.initialize(size));
PANDO_CHECK(forumevent_hastopic_topic_jihad.initialize(size));
PANDO_CHECK(publication_hasorg_topic_near_nyc.initialize(size));
PANDO_CHECK(publication_hastopic_topic_electrical_eng.initialize(size));

PANDO_CHECK(similarity.initialize(size));
for (pando::GlobalRef<pando::Vector<node_sim>> s : similarity) {
pando::Vector<node_sim> empty_s;
PANDO_CHECK(empty_s.initialize(other_size));
s = empty_s;
}
PANDO_CHECK(new_matched.initialize(size));
PANDO_CHECK(matched.initialize(size));
PANDO_CHECK(match.initialize(size));
}

void deinitialize() {
person_sale_person_bomb_bath.deinitialize();
person_sale_person_pressure_cooker.deinitialize();
person_sale_person_ammunition.deinitialize();
person_sale_person_electronics.deinitialize();
person_purchase_person_bomb_bath.deinitialize();
person_purchase_person_pressure_cooker.deinitialize();
person_purchase_person_ammunition.deinitialize();
person_purchase_person_electronics.deinitialize();
person_author_forumevent.deinitialize();
person_author_publication.deinitialize();
forum_includes_forumevent.deinitialize();
forum_hastopic_topic_nyc.deinitialize();
forumevent_hastopic_topic_bomb.deinitialize();
forumevent_hastopic_topic_explosion.deinitialize();
forumevent_hastopic_topic_williamsburg.deinitialize();
forumevent_hastopic_topic_outdoors.deinitialize();
forumevent_hastopic_topic_prospect_park.deinitialize();
forumevent_hastopic_topic_jihad.deinitialize();
publication_hasorg_topic_near_nyc.deinitialize();
publication_hastopic_topic_electrical_eng.deinitialize();
for (pando::Vector<node_sim> s : similarity) {
s.deinitialize();
}
similarity.deinitialize();
match.deinitialize();
matched.deinitialize();
new_matched.deinitialize();
}
};

struct State {
GraphState state_lhs;
GraphState state_rhs;
pando::GlobalPtr<std::uint64_t> match_count_ptr;

public:
State() = default;
State(Graph g_lhs, Graph g_rhs) : state_lhs(g_lhs), state_rhs(g_rhs) {}

void initialize() {
state_lhs.initialize(state_rhs.graph.size());
state_rhs.initialize(state_lhs.graph.size());
match_count_ptr = static_cast<decltype(match_count_ptr)>(
pando::getDefaultMainMemoryResource()->allocate(sizeof(std::uint64_t)));
}

void deinitialize() {
state_lhs.deinitialize();
state_rhs.deinitialize();
}
};

void match(pando::GlobalPtr<Graph>, pando::GlobalPtr<Graph>, int k);

void matchTriples(GraphState, Graph::VertexTopologyID);
void matchTriplesPerson(GraphState, Graph::VertexTopologyID);
void matchTriplesForum(GraphState, Graph::VertexTopologyID);
void matchTriplesForumEvent(GraphState, Graph::VertexTopologyID);
void matchTriplesPub(GraphState, Graph::VertexTopologyID);
void calculateSimilarity(State&);
void calculateMatch(State);

} // namespace approx
} // namespace wf2

#endif // PANDO_WF2_GALOIS_APPROX_MATCH_H_
Loading

0 comments on commit 9d88802

Please sign in to comment.