diff --git a/benchmark/word_size.cpp b/benchmark/word_size.cpp new file mode 100644 index 00000000..d054efcd --- /dev/null +++ b/benchmark/word_size.cpp @@ -0,0 +1,197 @@ +// Copyright 2021, 2022 Peter Dimov. +// Distributed under the Boost Software License, Version 1.0. +// https://www.boost.org/LICENSE_1_0.txt + +#define _SILENCE_CXX17_OLD_ALLOCATOR_MEMBERS_DEPRECATION_WARNING + +#include +#include +#include +#ifdef HAVE_ABSEIL +# include "absl/container/node_hash_map.h" +# include "absl/container/flat_hash_map.h" +#endif +#include +#include +#include +#include +#include +#include +#include +#include + +using namespace std::chrono_literals; + +static void print_time( std::chrono::steady_clock::time_point & t1, char const* label, std::size_t s, std::size_t size ) +{ + auto t2 = std::chrono::steady_clock::now(); + + std::cout << label << ": " << ( t2 - t1 ) / 1ms << " ms (s=" << s << ", size=" << size << ")\n"; + + t1 = t2; +} + +static std::vector words; + +static void init_words() +{ + char const* fn = "enwik9"; // http://mattmahoney.net/dc/textdata + + auto t1 = std::chrono::steady_clock::now(); + + std::ifstream is( fn ); + std::string in( std::istreambuf_iterator( is ), std::istreambuf_iterator{} ); + + boost::regex re( "[a-zA-Z]+"); + boost::sregex_token_iterator it( in.begin(), in.end(), re, 0 ), end; + + words.assign( it, end ); + + auto t2 = std::chrono::steady_clock::now(); + + std::cout << fn << ": " << words.size() << " words, " << ( t2 - t1 ) / 1ms << " ms\n\n"; +} + +template BOOST_NOINLINE void test_word_size( Map& map, std::chrono::steady_clock::time_point & t1 ) +{ + for( auto const& word: words ) + { + ++map[ word.size() ]; + } + + print_time( t1, "Word size count", 0, map.size() ); + + std::cout << std::endl; +} + +// counting allocator + +static std::size_t s_alloc_bytes = 0; +static std::size_t s_alloc_count = 0; + +template struct allocator +{ + using value_type = T; + + allocator() = default; + + template allocator( allocator const & ) noexcept + { + } + + template bool operator==( allocator const & ) const noexcept + { + return true; + } + + template bool operator!=( allocator const& ) const noexcept + { + return false; + } + + T* allocate( std::size_t n ) const + { + s_alloc_bytes += n * sizeof(T); + s_alloc_count++; + + return std::allocator().allocate( n ); + } + + void deallocate( T* p, std::size_t n ) const noexcept + { + s_alloc_bytes -= n * sizeof(T); + s_alloc_count--; + + std::allocator().deallocate( p, n ); + } +}; + +// + +struct record +{ + std::string label_; + long long time_; + std::size_t bytes_; + std::size_t count_; +}; + +static std::vector times; + +template class Map> BOOST_NOINLINE void test( char const* label ) +{ + std::cout << label << ":\n\n"; + + s_alloc_bytes = 0; + s_alloc_count = 0; + + Map map; + + auto t0 = std::chrono::steady_clock::now(); + auto t1 = t0; + + test_word_size( map, t1 ); + + std::cout << "Memory: " << s_alloc_bytes << " bytes in " << s_alloc_count << " allocations\n\n"; + + record rec = { label, 0, s_alloc_bytes, s_alloc_count }; + + auto tN = std::chrono::steady_clock::now(); + std::cout << "Total: " << ( tN - t0 ) / 1ms << " ms\n\n"; + + rec.time_ = ( tN - t0 ) / 1ms; + times.push_back( rec ); +} + +// aliases using the counting allocator + +template using allocator_for = ::allocator< std::pair >; + +template using std_unordered_map = + std::unordered_map, std::equal_to, allocator_for>; + +template using boost_unordered_map = + boost::unordered_map, std::equal_to, allocator_for>; + +template using boost_unordered_flat_map = + boost::unordered_flat_map, std::equal_to, allocator_for>; + +#ifdef HAVE_ABSEIL + +template using absl_node_hash_map = + absl::node_hash_map, absl::container_internal::hash_default_eq, allocator_for>; + +template using absl_flat_hash_map = + absl::flat_hash_map, absl::container_internal::hash_default_eq, allocator_for>; + +#endif + +int main() +{ + init_words(); + + test( "std::unordered_map" ); + test( "boost::unordered_map" ); + test( "boost::unordered_flat_map" ); + +#ifdef HAVE_ABSEIL + + test( "absl::node_hash_map" ); + test( "absl::flat_hash_map" ); + +#endif + + std::cout << "---\n\n"; + + for( auto const& x: times ) + { + std::cout << std::setw( 27 ) << ( x.label_ + ": " ) << std::setw( 5 ) << x.time_ << " ms, " << std::setw( 9 ) << x.bytes_ << " bytes in " << x.count_ << " allocations\n"; + } +} + +#ifdef HAVE_ABSEIL +# include "absl/container/internal/raw_hash_set.cc" +# include "absl/hash/internal/hash.cc" +# include "absl/hash/internal/low_level_hash.cc" +# include "absl/hash/internal/city.cc" +#endif