Skip to content
Tutorials Center
My learning

Topics

C++ Reference ยท Memory and Utilities

The <cmath> header

Reference for the C++ cmath header: sqrt, pow, trig, exp, log, floor, ceil, round, fmod, hypot, lerp, isnan, isfinite and the C++17 special functions.

<cmath> declares the standard math functions for float, double and long double, in namespace std (and usually also in the global namespace). It covers powers and roots, exponentials and logarithms, trigonometry, rounding, remainders, classification of special values, and a set of special functions such as the error function and Bessel functions.

The overloads pick the precision from the argument: std::sqrt(2.0f) returns a float. Integer arguments are converted to double, so std::sqrt(2) works. The constants pi and e are not in this header; they live in <numbers> (C++20).

C++
#include <cmath>
double h = std::hypot(3.0, 4.0);   // 5

Try it

C++ Rounding, roots and powers

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with std::cin or std::getline), it gets the next line from this box.

C++ Special values and classification

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with std::cin or std::getline), it gets the next line from this box.

Syntax

C++
double sqrt(double x);   float sqrt(float x);   long double sqrt(long double x);
double pow(double base, double exponent);
double fmod(double x, double y);
bool isnan(double x);    bool isfinite(double x);   bool isinf(double x);
Group Functions
Powers sqrt, cbrt, pow, hypot (two and, since C++17, three arguments)
Exponential exp, exp2, expm1, log, log2, log10, log1p
Trigonometry sin, cos, tan, asin, acos, atan, atan2 and the hyperbolic versions sinh, cosh, tanh, asinh, acosh, atanh (radians, not degrees)
Rounding floor, ceil, trunc, round, lround, llround, rint, nearbyint
Remainder fmod, remainder, remquo
Absolute value, sign abs, fabs, copysign, signbit
Comparison fmin, fmax, fdim, isless, isgreater and friends
Classification isnan, isinf, isfinite, isnormal, fpclassify
Special (C++17) erf, erfc, tgamma, lgamma, beta, comp_ellint_1, cyl_bessel_j, riemann_zeta and others
Other lerp (C++20), fma, ldexp, frexp, modf, nextafter

Errors: domain and range errors are reported through errno and the floating-point exception flags, depending on math_errhandling; by default with GCC the functions just return NaN or infinity and set errno (EDOM, ERANGE). They do not throw.

Usage notes

  • Angles are radians. Convert with degrees * std::numbers::pi / 180.
  • Use std::abs, not C's abs, for floating-point values: the plain abs(int) from <cstdlib> would truncate a double.
  • std::pow with an integer exponent is a floating-point call. For exact integer powers write a loop (or 1 << n for powers of two); a double result is exact only while it fits in 53 bits.
  • Compare floats with a tolerance, and never with == after arithmetic. NaN is never equal to anything, including itself.
  • round vs rint: round always moves ties away from zero; rint follows the current rounding mode, which is to-nearest-even by default.
  • -ffast-math breaks NaN and infinity handling. Avoid it when you call isnan.
  • constexpr: most <cmath> functions are not constexpr before C++26, although GCC folds many of them at compile time as an extension.
  • Prefer <cmath> to <math.h>. It gives you the overload set; the C header is only guaranteed for the double versions.