Search⌘ K
AI Features
Home
Courses
C++ High Performance

Template Creation

Explore how to create C++ templates with integral parameters and specialize them for specific cases. Understand how the compiler generates functions from templates, use abbreviated function templates introduced in C++20, and employ decltype and std::remove_cvref for type manipulation. This lesson enhances your ability to write efficient, generic compile-time code.

We'll cover the following...

Using integers as template parameters

Beyond general types, a template can also be of other types, such as integral types and floating-point types. In the following example, we will use an int in the template, which means that the compiler will generate a new function for every unique integer passed as a template argument:

C++
template <int N, typename T> 
auto const_pow_n(const T& v) {
  auto product = T{1};
  for (int i = 0; i < N; ++i) {
    product *= v;
  }
  return product;
}

The following code will oblige the compiler to instantiate two distinct functions: one squares the value and one cubes the value:

C++
auto x2 = const_pow_n<2>(4.0f); // Square 
auto x3 = const_pow_n<3>(4.0f); // Cube

Note the difference between the template parameter N and the function parameter v. For every value of N, the compiler generates a new function. However, v is passed as a regular parameter and, as such, does not result in a new function.

Providing specializations of a template

By default, the compiler will generate regular C++ code whenever we use a template with new parameters. But it’s also possible to provide a custom implementation for certain values of the template parameters. Say, for example, that we want to provide the regular C++ code of our const_pow_n() function when it’s used with integers and the value of ...