嵌套的 std::transform 效率低下吗？

Question

如果我有 std::string:

std::string s{"hello"};

和一个就地修改它的循环，像这样：

for (auto &c: s)
  c = std::toupper(c);

我可以用等效的 transform:

替换它

std::transform(s.begin(), s.end(), s.begin(),
               [](unsigned char c) -> unsigned char 
               { return std::toupper(c); });

这些片段生成相同的 assembly. They also have similar performance。

但是，如果我有 std::vector<std::string>:

std::vector<std::string> v {"hello", "how", "are", "you"};

并像这样就地修改它：

for (auto & s : v)
  for (auto &c: s)
    c = std::toupper(c);

等效变换应该是：

std::transform(std::begin(v), std::end(v), std::begin(v),
  [](auto s) {
    std::transform(std::begin(s), std::end(s), std::begin(s), 
      [](unsigned char c) -> unsigned char { return std::toupper(c); });
    return s;
});

然而，transform 版本产生了一半以上 assembly, and performs 相应地很差，这让我感到惊讶。

std::transform在这种情况下不是零成本抽象，还是我只是使用不当？

Answer 1

通过并return 引用所有内容。否则，您会制作该字符串的多个副本。注意变化：[](auto& s) -> std::string& {

std::transform(std::begin(v), std::end(v), std::begin(v),
  [](auto& s) -> std::string& {
    std::transform(std::begin(s), std::end(s), std::begin(s), 
      [](unsigned char c) -> unsigned char { return std::toupper(c); });
    return s;
});

我向您的 link 添加了两个新的快速工作台功能。一种将输入字符串作为引用传递。另一个也是 return 的引用。即：

static void Transform2(benchmark::State& state) {
  // Code before the loop is not measured

  std::vector<std::string> v {"hello", "how", "are", "you"};
  for (auto _ : state) {
    std::transform(std::begin(v), std::end(v), std::begin(v),
    [](auto& s) {
      std::transform(std::begin(s), std::end(s), std::begin(s), 
        [](unsigned char c) -> unsigned char { return std::toupper(c); });
      return s;
    });

  }
}
BENCHMARK(Transform2);


static void Transform3(benchmark::State& state) {
  // Code before the loop is not measured

  std::vector<std::string> v {"hello", "how", "are", "you"};
  for (auto _ : state) {
    std::transform(std::begin(v), std::end(v), std::begin(v),
    [](auto& s) -> std::string& {
      std::transform(std::begin(s), std::end(s), std::begin(s), 
        [](unsigned char c) -> unsigned char { return std::toupper(c); });
      return s;
    });

  }
}
BENCHMARK(Transform3);

取决于我在运行基准测试时的幸运程度，Transform3 的性能几乎（有时等于）InPlace 测试实现。

嵌套的 std::transform 效率低下吗？

Is a nested std::transform inefficient?

c++

performance

stl-algorithm