ceph/src/arrow/cpp/src/arrow/util/utf8_util_benchmark.cc

   1 // Licensed to the Apache Software Foundation (ASF) under one
   2 // or more contributor license agreements.  See the NOTICE file
   3 // distributed with this work for additional information
   4 // regarding copyright ownership.  The ASF licenses this file
   5 // to you under the Apache License, Version 2.0 (the
   6 // "License"); you may not use this file except in compliance
   7 // with the License.  You may obtain a copy of the License at
   8 //
   9 //   http://www.apache.org/licenses/LICENSE-2.0
  10 //
  11 // Unless required by applicable law or agreed to in writing,
  12 // software distributed under the License is distributed on an
  13 // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
  14 // KIND, either express or implied.  See the License for the
  15 // specific language governing permissions and limitations
  16 // under the License.
  17
  18 #include "benchmark/benchmark.h"
  19
  20 #include <string>
  21 #include <type_traits>
  22 #include <vector>
  23
  24 #include "arrow/testing/gtest_util.h"
  25 #include "arrow/util/utf8.h"
  26
  27 // Do not benchmark inlined functions directly inside the benchmark loop
  28 static ARROW_NOINLINE bool ValidateUTF8NoInline(const uint8_t* data, int64_t size) {
  29   return ::arrow::util::ValidateUTF8(data, size);
  30 }
  31
  32 static ARROW_NOINLINE bool ValidateAsciiNoInline(const uint8_t* data, int64_t size) {
  33   return ::arrow::util::ValidateAscii(data, size);
  34 }
  35
  36 namespace arrow {
  37 namespace util {
  38
  39 static const char* tiny_valid_ascii = "characters";
  40 static const char* tiny_valid_non_ascii = "caractères";
  41
  42 static const char* valid_ascii =
  43     "UTF-8 is a variable width character encoding capable of encoding all 1,112,064 "
  44     "valid code points in Unicode using one to four 8-bit bytes";
  45 static const char* valid_almost_ascii =
  46     "UTF-8 est un codage de caractères informatiques conçu pour coder l’ensemble des "
  47     "caractères du « répertoire universel de caractères codés »";
  48 static const char* valid_non_ascii =
  49     "UTF-8 はISO/IEC 10646 (UCS) "
  50     "とUnicodeで使える8ビット符号単位の文字符号化形式及び文字符号化スキーム。 ";
  51
  52 static std::string MakeLargeString(const std::string& base, int64_t nbytes) {
  53   int64_t nrepeats = (nbytes + base.size() - 1) / base.size();
  54   std::string s;
  55   s.reserve(nrepeats * nbytes);
  56   for (int64_t i = 0; i < nrepeats; ++i) {
  57     s += base;
  58   }
  59   return s;
  60 }
  61
  62 static void BenchmarkUTF8Validation(
  63     benchmark::State& state,  // NOLINT non-const reference
  64     const std::string& s, bool expected) {
  65   auto data = reinterpret_cast<const uint8_t*>(s.data());
  66   auto data_size = static_cast<int64_t>(s.size());
  67
  68   InitializeUTF8();
  69   bool b = ValidateUTF8NoInline(data, data_size);
  70   if (b != expected) {
  71     std::cerr << "Unexpected validation result" << std::endl;
  72     std::abort();
  73   }
  74
  75   while (state.KeepRunning()) {
  76     bool b = ValidateUTF8NoInline(data, data_size);
  77     benchmark::DoNotOptimize(b);
  78   }
  79   state.SetBytesProcessed(state.iterations() * s.size());
  80 }
  81
  82 static void BenchmarkASCIIValidation(
  83     benchmark::State& state,  // NOLINT non-const reference
  84     const std::string& s, bool expected) {
  85   auto data = reinterpret_cast<const uint8_t*>(s.data());
  86   auto data_size = static_cast<int64_t>(s.size());
  87
  88   bool b = ValidateAsciiNoInline(data, data_size);
  89   if (b != expected) {
  90     std::cerr << "Unexpected validation result" << std::endl;
  91     std::abort();
  92   }
  93
  94   while (state.KeepRunning()) {
  95     bool b = ValidateAsciiNoInline(data, data_size);
  96     benchmark::DoNotOptimize(b);
  97   }
  98   state.SetBytesProcessed(state.iterations() * s.size());
  99 }
 100
 101 static void ValidateTinyAscii(benchmark::State& state) {  // NOLINT non-const reference
 102   BenchmarkASCIIValidation(state, tiny_valid_ascii, true);
 103 }
 104
 105 static void ValidateTinyNonAscii(benchmark::State& state) {  // NOLINT non-const reference
 106   BenchmarkUTF8Validation(state, tiny_valid_non_ascii, true);
 107 }
 108
 109 static void ValidateSmallAscii(benchmark::State& state) {  // NOLINT non-const reference
 110   BenchmarkASCIIValidation(state, valid_ascii, true);
 111 }
 112
 113 static void ValidateSmallAlmostAscii(
 114     benchmark::State& state) {  // NOLINT non-const reference
 115   BenchmarkUTF8Validation(state, valid_almost_ascii, true);
 116 }
 117
 118 static void ValidateSmallNonAscii(
 119     benchmark::State& state) {  // NOLINT non-const reference
 120   BenchmarkUTF8Validation(state, valid_non_ascii, true);
 121 }
 122
 123 static void ValidateLargeAscii(benchmark::State& state) {  // NOLINT non-const reference
 124   auto s = MakeLargeString(valid_ascii, 100000);
 125   BenchmarkASCIIValidation(state, s, true);
 126 }
 127
 128 static void ValidateLargeAlmostAscii(
 129     benchmark::State& state) {  // NOLINT non-const reference
 130   auto s = MakeLargeString(valid_almost_ascii, 100000);
 131   BenchmarkUTF8Validation(state, s, true);
 132 }
 133
 134 static void ValidateLargeNonAscii(
 135     benchmark::State& state) {  // NOLINT non-const reference
 136   auto s = MakeLargeString(valid_non_ascii, 100000);
 137   BenchmarkUTF8Validation(state, s, true);
 138 }
 139
 140 BENCHMARK(ValidateTinyAscii);
 141 BENCHMARK(ValidateTinyNonAscii);
 142 BENCHMARK(ValidateSmallAscii);
 143 BENCHMARK(ValidateSmallAlmostAscii);
 144 BENCHMARK(ValidateSmallNonAscii);
 145 BENCHMARK(ValidateLargeAscii);
 146 BENCHMARK(ValidateLargeAlmostAscii);
 147 BENCHMARK(ValidateLargeNonAscii);
 148
 149 }  // namespace util
 150 }  // namespace arrow