# std::cos, std::cosf, std::cosl

> C++ Atlas: `cppatlas.com/w/cpp/numeric/math/cos` | Source: https://en.cppreference.com/w/cpp/numeric/math/cos | Domain: Numerics | Standards: C++11, C++23, C++26

If no errors occur, the cosine of num (cos(num)) in the range [-1.0, +1.0], is returned.

## Official Specification Summary (cppreference)

```cpp
// Defined in header <cmath>

float       cos ( float num );
double      cos ( double num );
long double cos ( long double num ); // (1) (until C++23)

/*floating-point-type*/
            cos ( /*floating-point-type*/ num ); // (1) (since C++23) (constexpr since C++26)

float       cosf( float num ); // (2) (since C++11) (constexpr since C++26)

long double cosl( long double num ); // (3) (since C++11) (constexpr since C++26)

// SIMD overload (since C++26)

// Defined in header <simd>

template< /*math-floating-point*/ V >
constexpr /*deduced-simd-t*/<V>
            cos ( const V& v_num ); // (S) (since C++26)

// Additional overloads (since C++11)

// Defined in header <cmath>

template< class Integer >
double      cos ( Integer num ); // (A) (constexpr since C++26)
```

1-3) Computes the cosine of `num` (measured in radians). The library provides overloads of `std::cos` for all cv-unqualified floating-point types as the type of the parameter. (since C++23)

(since C++26):

S) The SIMD overload performs an element-wise `std::cos` on `v_num`.

  (See `math-floating-point` and `deduced-simd-t` for their definitions.)

A) Additional overloads are provided for all integer types, which are treated as `double`. (since C++11)

### Parameters

- `num` - floating-point or integer value representing angle in radians

### Return value

If no errors occur, the cosine of `num` (cos(num)) in the range `[-1.0,` `+1.0]`, is returned.

The result may have little or no significance if the magnitude of `num` is large. (until C++11)

If a domain error occurs, an implementation-defined value is returned (NaN where supported).

If a range error occurs due to underflow, the correct result (after rounding) is returned.

### Error handling

Errors are reported as specified in `math_errhandling`.

If the implementation supports IEEE floating-point arithmetic (IEC 60559),

- if the argument is ±0, the result is `1.0`.
- if the argument is ±∞, NaN is returned and `FE_INVALID` is raised.
- if the argument is NaN, NaN is returned.

### Notes

The case where the argument is infinite is not specified to be a domain error in C, but it is defined as a domain error in POSIX.

The additional overloads are not required to be provided exactly as (A). They only need to be sufficient to ensure that for their argument `num` of integer type, `std::cos(num)` has the same effect as `std::cos(static_cast<double>(num))`.

### Example

```cpp
#include <cerrno>
#include <cfenv>
#include <cmath>
#include <iomanip>
#include <iostream>
#include <numbers>

// #pragma STDC FENV_ACCESS ON

constexpr double pi = std::numbers::pi; // or std::acos(-1) before C++20

constexpr double your_cos(double x)
{
    double cos{1}, pow{x};
    for (auto fac{1ull}, n{1ull}; n != 19; fac *= ++n, pow *= x)
        if ((n & 1) == 0)
            cos += (n & 2 ? -pow : pow) / fac;
    return cos;
}

int main()
{
    std::cout << std::setprecision(10) << std::showpos
              << "Typical usage:\n"
              << "std::cos(pi/3) = " << std::cos(pi / 3) << '\n'
              << "your cos(pi/3) = " << your_cos(pi / 3) << '\n'
              << "std::cos(pi/2) = " << std::cos(pi / 2) << '\n'
              << "your cos(pi/2) = " << your_cos(pi / 2) << '\n'
              << "std::cos(-3*pi/4) = " << std::cos(-3 * pi / 4) << '\n'
              << "your cos(-3*pi/4) = " << your_cos(-3 * pi / 4) << '\n'
              << "Special values:\n"
              << "std::cos(+0) = " << std::cos(0.0) << '\n'
              << "std::cos(-0) = " << std::cos(-0.0) << '\n';

    // error handling
    std::feclearexcept(FE_ALL_EXCEPT);

    std::cout << "cos(INFINITY) = " << std::cos(INFINITY) << '\n';
    if (std::fetestexcept(FE_INVALID))
        std::cout << "    FE_INVALID raised\n";
}
```

Possible output:

```text
Typical usage:
std::cos(pi/3) = +0.5
your cos(pi/3) = +0.5
std::cos(pi/2) = +6.123233996e-17
your cos(pi/2) = -3.373452105e-15
std::cos(-3*pi/4) = -0.7071067812
your cos(-3*pi/4) = -0.7071067812
Special values:
std::cos(+0) = +1
std::cos(-0) = +1
cos(INFINITY) = -nan
    FE_INVALID raised
```

### See also

- [`sin`](https://en.cppreference.com/w/cpp/numeric/math/sin), [`sinf`](https://en.cppreference.com/w/cpp/numeric/math/sin) (C++11), [`sinl`](https://en.cppreference.com/w/cpp/numeric/math/sin) (C++11): computes sine (sin(x)) (function)
- [`tan`](https://en.cppreference.com/w/cpp/numeric/math/tan), [`tanf`](https://en.cppreference.com/w/cpp/numeric/math/tan) (C++11), [`tanl`](https://en.cppreference.com/w/cpp/numeric/math/tan) (C++11): computes tangent (tan(x)) (function)
- [`acos`](https://en.cppreference.com/w/cpp/numeric/math/acos), [`acosf`](https://en.cppreference.com/w/cpp/numeric/math/acos) (C++11), [`acosl`](https://en.cppreference.com/w/cpp/numeric/math/acos) (C++11): computes arc cosine (arccos(x)) (function)
- [`cos(std::complex)`](https://en.cppreference.com/w/cpp/numeric/complex/cos): computes cosine of a complex number (cos(z)) (function template)
- [`cos(std::valarray)`](https://en.cppreference.com/w/cpp/numeric/valarray/cos): applies the function `std::cos` to each element of valarray (function template)
- [C documentation for `cos`](https://en.cppreference.com/w/c/numeric/math/cos)

---
Attribution: content from https://en.cppreference.com/w/cpp/numeric/math/cos (cppreference.com), licensed under CC-BY-SA 3.0 and GFDL. Modified: converted to markdown, study notes added. Offline snapshot of 9 February 2025.