clean up cmake setup

BlaiseMuhirwa · Jan 14, 2024 · 18439cb · 18439cb
1 parent 30f1990
commit 18439cb
Show file tree

Hide file tree

Showing 13 changed files with 254 additions and 275 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -14,8 +14,6 @@ include_directories(${CMAKE_SOURCE_DIR})
 # Include cereal
 include_directories(${CMAKE_SOURCE_DIR}/external/cereal/include/)
 
-# set standard to c++17 in order to use std::unique_ptr(c++14) and nested
-# namespaces (c++17)
 set(CMAKE_CXX_STANDARD 17)
 
 set(CMAKE_CXX_FLAGS
@@ -29,159 +27,23 @@ set(CMAKE_CXX_FLAGS
     -ffast-math \
     -funroll-loops")
 
-# Code adapted from PyTorch:
-# https://github.com/pytorch/pytorch/blob/main/cmake/Modules/FindAVX.cmake
-include(CheckCXXSourceRuns)
-set(AVX_CODE
-    "
-      #include <immintrin.h>
-      int main() {
-        __m256 a;
-        a = _mm256_set1_ps(0);
-        return 0;
-      }
-    ")
-
-set(AVX512_CODE
-    "
-    #include <immintrin.h>
-    int main() {
-      __m512 a = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0,
-                                  0, 0, 0, 0, 0, 0, 0, 0);
-      __m512i b = a;
-      __mmask64 equality_mask = _mm512_cmp_epi8_mask(a, b, _MM_CMPINT_EQ);
-      return 0;
-      )
-    }
-    ")
-
-include(CheckCXXCompilerFlag)
-
-# Function to check compiler support and hardware capability for a given flag
-function(check_compiler_and_hardware_support FLAG CODE_VAR EXTENSION_NAME)
-  check_cxx_compiler_flag(${FLAG} COMPILER_SUPPORTS_${EXTENSION_NAME})
-  if(COMPILER_SUPPORTS_${EXTENSION_NAME})
-    set(CMAKE_REQUIRED_FLAGS_SAVE ${CMAKE_REQUIRED_FLAGS})
-    set(CMAKE_REQUIRED_FLAGS "${CMAKE_REQUIRED_FLAGS} ${FLAG}")
-
-    check_cxx_source_runs("${${CODE_VAR}}"
-                          SYSTEM_SUPPORTS_${EXTENSION_NAME}_EXTENSIONS)
-    set(CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS_SAVE})
-
-    if(SYSTEM_SUPPORTS_${EXTENSION_NAME}_EXTENSIONS)
-      set(CMAKE_CXX_FLAGS
-          "${CMAKE_CXX_FLAGS} ${FLAG}"
-          PARENT_SCOPE)
-      message(STATUS "Building with ${EXTENSION_NAME}")
-    else()
-      message(
-        STATUS "Compiler supports ${FLAG} flag but the target machine does not "
-               "support ${EXTENSION_NAME} instructions")
-    endif()
-  endif()
-endfunction()
-
-# Build SSE/AVX/AVX512 code only on x86-64 processors.
-if(CMAKE_HOST_SYSTEM_PROCESSOR MATCHES "(x86_64)|(AMD64|amd64)|(^i.86$)")
-  check_compiler_and_hardware_support("-mavx512f" "AVX512_CODE" "AVX512")
-  check_compiler_and_hardware_support("-mavx" "AVX_CODE" "AVX")
-
-  check_cxx_compiler_flag("-msse" CXX_SSE)
-  if(CXX_SSE)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse")
-    message(STATUS "Building with SSE")
-  endif()
-endif()
+list(APPEND CMAKE_MODULE_PATH "${PROJECT_SOURCE_DIR}/cmake")
+include(cmake/FindAVX.cmake)
 
 option(CMAKE_BUILD_TYPE "Build type" Release)
 if(CMAKE_BUILD_TYPE STREQUAL "Debug")
   # Add debug compile flags
   message(STATUS "Building in Debug mode")
-  # For more on address sanitizer:
-  # https://clang.llvm.org/docs/AddressSanitizer.html ASan is supposed to be
-  # very fast, but if we find it slow, we can remove it or use compiler
-  # directives to skip analyzing functions: __attribute__((no_sanitize_address))
-  # add_compile_options(-g -Wall -fsanitize=address)
+  # Address sanitizer: https://clang.llvm.org/docs/AddressSanitizer.html
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -g -Wall -fsanitize=address")
 endif()
 
-include(ExternalProject)
 include(FeatureSummary)
-include(FetchContent)
-
-find_package(Git REQUIRED)
-
-option(USE_GIT_PROTOCOL
-       "If behind a firewall turn this off to use HTTPS instead." OFF)
-
 # All options summary
 feature_summary(WHAT ALL)
 
-# Configure options
-configure_file("${PROJECT_SOURCE_DIR}/quantization/Config.h.in"
-               "${PROJECT_BINARY_DIR}/Config.h")
 include_directories("${PROJECT_BINARY_DIR}" "${PROJECT_SOURCE_DIR}")
 
-function(functionInstallExternalCMakeProject ep_name)
-  ExternalProject_Get_Property(${ep_name} binary_dir)
-  install(SCRIPT ${binary_dir}/cmake_install.cmake)
-endfunction()
-
-# TODO(blaise): Remove this old school way of including external libraries via
-# ExternalProject_Add. Use FetchContent instead, which is much cleaner
-ExternalProject_Add(
-  ZLIB
-  DEPENDS ""
-  GIT_REPOSITORY https://github.com/madler/zlib.git
-  GIT_TAG v1.2.11
-  SOURCE_DIR ZLIB-source
-  BINARY_DIR ZLIB-build
-  UPDATE_COMMAND ""
-  PATCH_COMMAND ""
-  # INSTALL_COMMAND ""
-  CMAKE_GENERATOR ${gen}
-  CMAKE_ARGS
-    -DCMAKE_INSTALL_PREFIX:STRING=${PROJECT_BINARY_DIR}/ep
-    -DINSTALL_BIN_DIR:STRING=${PROJECT_BINARY_DIR}/ep/bin
-    -DINSTALL_INC_DIR:STRING=${PROJECT_BINARY_DIR}/ep/include
-    -DINSTALL_LIB_DIR:STRING=${PROJECT_BINARY_DIR}/ep/lib
-    -DINSTALL_MAN_DIR:STRING=${PROJECT_BINARY_DIR}/ep/share/man
-    -DINSTALL_PKGCONFIG_DIR:STRING=${PROJECT_BINARY_DIR}/ep/share/pkgconfig
-    -DCMAKE_BUILD_TYPE:STRING=Release)
-functioninstallexternalcmakeproject(ZLIB)
-
-set(ZLIB_LIB_DEBUG ${PROJECT_BINARY_DIR}/ep/lib/libz.a)
-set(ZLIB_LIB_RELEASE ${PROJECT_BINARY_DIR}/ep/lib/libz.a)
-
-ExternalProject_Add(
-  CNPY
-  DEPENDS ZLIB
-  GIT_REPOSITORY https://github.com/sarthakpati/cnpy.git
-  # GIT_TAG v1.2.11
-  SOURCE_DIR CNPY-source
-  BINARY_DIR CNPY-build
-  UPDATE_COMMAND ""
-  PATCH_COMMAND ""
-  # INSTALL_COMMAND ""
-  CMAKE_GENERATOR ${gen}
-  CMAKE_ARGS -DZLIB_INCLUDE_DIR:STRING=${PROJECT_BINARY_DIR}/ep/include
-             -DZLIB_LIBRARY_DEBUG:STRING=${ZLIB_LIB_DEBUG}
-             -DZLIB_LIBRARY_RELEASE:STRING=${ZLIB_LIB_RELEASE}
-             -DCMAKE_INSTALL_PREFIX:STRING=${PROJECT_BINARY_DIR}/ep
-             -DBUILD_SHARED_LIBS:BOOL=${BUILD_SHARED_LIBS}
-             -DCMAKE_BUILD_TYPE:STRING=Release)
-functioninstallexternalcmakeproject(CNPY)
-
-include_directories(${PROJECT_BINARY_DIR}/ep/include)
-
-set(CNPY_LIB ${PROJECT_BINARY_DIR}/ep/lib/libcnpy.a)
-
 find_package(OpenMP REQUIRED)
 if(OpenMP_FOUND)
   message(STATUS "OpenMP Found. Building the Package using the system OpenMP.")
@@ -207,50 +69,6 @@ if(NO_MANUAL_VECTORIZATION)
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -ftree-vectorize")
 endif()
 
-if(BUILD_BENCHMARKS)
-  enable_testing()
-
-  message(STATUS "Running a suite of benchmarks using Google Benchmark")
-  set(GOOGLE_BENCHMARK_DIR "${PROJECT_BINARY_DIR}/include/google-benchmark")
-
-  if(NOT EXISTS ${GOOGLE_BENCHMARK_DIR})
-    message(STATUS "Downloading google-benchmark to ${GOOGLE_BENCHMARK_DIR}")
-
-    FetchContent_Declare(
-      googletest
-      GIT_REPOSITORY https://github.com/google/googletest.git
-      GIT_TAG main)
-    FetchContent_Declare(
-      googlebenchmark
-      GIT_REPOSITORY https://github.com/google/benchmark.git
-      GIT_TAG main) # need main for benchmark::benchmark
-
-    FetchContent_MakeAvailable(googletest googlebenchmark)
-
-  else()
-    add_subdirectory(${GOOGLE_BENCHMARK_DIR})
-  endif()
-
-  set(YAML_CPP_DIR "${PROJECT_SOURCE_DIR}/external/yaml-cpp")
-
-  # Disable building tests for YAML-CPP
-  set(YAML_CPP_BUILD_TESTS
-      OFF
-      CACHE BOOL "Enable testing" FORCE)
-  # Fetch the YAML submodule if it doesn't exist
-  if(NOT EXISTS ${YAML_CPP_DIR})
-    message(STATUS "Fetching yaml-cpp submodule")
-    FetchContent_Declare(
-      yamlcpp
-      GIT_REPOSITORY https://github.com/jbeder/yaml-cpp.git
-      GIT_TAG yaml-cpp-0.6.3)
-    FetchContent_MakeAvailable(yamlcpp)
-  else()
-    add_subdirectory(${YAML_CPP_DIR})
-  endif()
-
-endif()
-
 # TODO: Using globbing or some other command that does not require writing down
 # every header file
 set(HEADERS
@@ -268,67 +86,31 @@ set(HEADERS
     ${PROJECT_SOURCE_DIR}/quantization/CentroidsGenerator.h
     ${PROJECT_SOURCE_DIR}/quantization/Utils.h)
 
-add_library(FLAT_NAV_LIB STATIC ${PROJECT_SOURCE_DIR}/tools/construct_npy.cpp
-                                ${HEADERS})
-add_dependencies(FLAT_NAV_LIB CNPY)
+# Interface here means that the library is header only
+add_library(FLAT_NAV_LIB INTERFACE)
+target_sources(FLAT_NAV_LIB INTERFACE ${HEADERS})
+target_include_directories(FLAT_NAV_LIB INTERFACE ${PROJECT_SOURCE_DIR})
 
-target_link_libraries(FLAT_NAV_LIB PUBLIC ${CNPY_LIB} OpenMP::OpenMP_CXX)
-set_target_properties(FLAT_NAV_LIB PROPERTIES LINKER_LANGUAGE CXX)
+target_link_libraries(FLAT_NAV_LIB INTERFACE OpenMP::OpenMP_CXX)
 
 if(BUILD_EXAMPLES)
-  message(STATUS "Building examples for Flatnav")
-  foreach(CONSTRUCT_EXEC construct_npy query_npy cereal_tests)
-    add_executable(${CONSTRUCT_EXEC}
-                   ${PROJECT_SOURCE_DIR}/tools/${CONSTRUCT_EXEC}.cpp ${HEADERS})
-    add_dependencies(${CONSTRUCT_EXEC} FLAT_NAV_LIB)
-    target_link_libraries(${CONSTRUCT_EXEC} FLAT_NAV_LIB ${CNPY_LIB}
-                          ${ZLIB_LIB_RELEASE})
-    install(TARGETS ${CONSTRUCT_EXEC} DESTINATION bin)
-  endforeach(CONSTRUCT_EXEC)
-
+  message(STATUS "Building examples")
+  list(APPEND CMAKE_MODULE_PATH "${PROJECT_SOURCE_DIR}/cmake")
+  include(cmake/FindCNPYAndZLIB.cmake)
+  add_subdirectory(${PROJECT_SOURCE_DIR}/tools)
 endif()
 
 if(BUILD_TESTS)
   message(STATUS "Building flatnav + quantization unit tests using gtest")
-  set(GOOGLE_TEST_DIR "${PROJECT_BINARY_DIR}/_deps/googletest-src")
-
-  if(NOT EXISTS ${GOOGLE_TEST_DIR})
-    message(
-      STATUS
-        "Downloading googletest to ${PROJECT_BINARY_DIR}/_deps/googletest-src")
-  endif()
-  # This does not download googletest again if its already available in the
-  # CMakeCache file
-  FetchContent_Declare(
-    googletest
-    GIT_REPOSITORY https://github.com/google/googletest.git
-    GIT_TAG main)
-
-  FetchContent_MakeAvailable(googletest)
-
-  FetchContent_GetProperties(googletest)
-  if(NOT googletest_POPULATED)
-    message(STATUS "GoogleTest not populated")
-    FetchContent_Populate(googletest)
-    add_subdirectory(${googletest_SOURCE_DIR} ${googletest_BINARY_DIR})
-  else()
-    message(STATUS "GoogleTest already populated")
-  endif()
-
-  message(STATUS "googletest_BINARY_DIR: ${googletest_BINARY_DIR}")
-  message(STATUS "googletest_SOURCE_DIR: ${googletest_SOURCE_DIR}")
-
-  target_include_directories(FLAT_NAV_LIB
-                             PUBLIC ${googletest_SOURCE_DIR}/googletest/include)
+  include(cmake/FindGoogleTest.cmake)
   add_subdirectory(${PROJECT_SOURCE_DIR}/flatnav/tests)
-  add_subdirectory(${PROJECT_SOURCE_DIR}/quantization/tests)
-
+  # add_subdirectory(${PROJECT_SOURCE_DIR}/quantization/tests)
 endif()
 
 if(BUILD_BENCHMARKS)
-  target_link_libraries(FLAT_NAV_LIB benchmark::benchmark)
-  add_executable(run_benchmark ${PROJECT_SOURCE_DIR}/benchmarks/runner.cpp)
-  target_link_libraries(run_benchmark benchmark::benchmark yaml-cpp
-                        FLAT_NAV_LIB ${CNPY_LIB} ${ZLIB_LIB_RELEASE})
-  install(TARGETS run_benchmark DESTINATION BIN)
+  message(STATUS "Building benchmarks")
+  if(NOT TARGET CNPY)
+    include(cmake/FindCNPYAndZLIB.cmake)
+  endif()
+  add_subdirectory(${PROJECT_SOURCE_DIR}/benchmarks)
 endif()
diff --git a/benchmarks/CMakeLists.txt b/benchmarks/CMakeLists.txt
@@ -0,0 +1,47 @@
+cmake_minimum_required(VERSION 3.14 FATAL_ERROR)
+enable_testing()
+
+include(FetchContent)
+message(STATUS "Running a suite of benchmarks using Google Benchmark")
+set(GOOGLE_BENCHMARK_DIR "${PROJECT_BINARY_DIR}/include/google-benchmark")
+
+if(NOT EXISTS ${GOOGLE_BENCHMARK_DIR})
+  message(STATUS "Downloading google-benchmark to ${GOOGLE_BENCHMARK_DIR}")
+
+  FetchContent_Declare(
+    googletest
+    GIT_REPOSITORY https://github.com/google/googletest.git
+    GIT_TAG main)
+  FetchContent_Declare(
+    googlebenchmark
+    GIT_REPOSITORY https://github.com/google/benchmark.git
+    GIT_TAG main) # need main for benchmark::benchmark
+  FetchContent_MakeAvailable(googletest googlebenchmark)
+
+else()
+  add_subdirectory(${GOOGLE_BENCHMARK_DIR})
+endif()
+
+set(YAML_CPP_DIR "${PROJECT_SOURCE_DIR}/external/yaml-cpp")
+
+# Disable building tests for YAML-CPP
+set(YAML_CPP_BUILD_TESTS
+    OFF
+    CACHE BOOL "Enable testing" FORCE)
+# Fetch the YAML submodule if it doesn't exist
+if(NOT EXISTS ${YAML_CPP_DIR})
+  message(STATUS "Fetching yaml-cpp submodule")
+  FetchContent_Declare(
+    yamlcpp
+    GIT_REPOSITORY https://github.com/jbeder/yaml-cpp.git
+    GIT_TAG yaml-cpp-0.6.3)
+  FetchContent_MakeAvailable(yamlcpp)
+else()
+  add_subdirectory(${YAML_CPP_DIR})
+endif()
+
+target_link_libraries(FLAT_NAV_LIB INTERFACE benchmark::benchmark)
+add_executable(run_benchmark ${PROJECT_SOURCE_DIR}/benchmarks/runner.cpp)
+target_link_libraries(run_benchmark benchmark::benchmark yaml-cpp
+                      FLAT_NAV_LIB ${CNPY_LIB} ${ZLIB_LIB_RELEASE})
+install(TARGETS run_benchmark DESTINATION BIN)
diff --git a/benchmarks/config_parser.h b/benchmarks/config_parser.h
@@ -5,7 +5,6 @@
 #include <stdint.h>
 #include <string>
 #include <string_view>
-#include <toml++/toml.h>
 #include <unordered_map>
 #include <vector>
 #include <yaml-cpp/yaml.h>

diff --git a/benchmarks/runner.cpp b/benchmarks/runner.cpp
@@ -32,11 +32,9 @@ std::unordered_map<uint32_t, std::string> dataset_id_to_filepath;
 template <typename dist_t>
 void buildIndex(std::shared_ptr<Index<dist_t, int>> index, float *data, int N,
                 int M, int dim, int ef_construction) {
-  for (int label = 0; label < N; label++) {
-    float *element = data + (dim * label);
-    index->add(/* data = */ (void *)element, /* label = */ label,
-               /* ef_construction */ ef_construction);
-  }
+  std::vector<int> labels(N);
+  std::iota(labels.begin(), labels.end(), 0);
+  index->addBatch(data, labels, ef_construction);
 }
 
 template <typename dist_t> struct BenchmarkFixture : public benchmark::Fixture {