Nira

Nira is a header-only C++20 library that provides constexpr arithmetic types that model different types of numbers. The types provided implement arithmetic operations like addition, subtraction, multiplication, division, and comparisons. They can be used as drop-in replacements for built-in arithmetic types like int or float.

nira::Rational

Rational models a rational number, a number which can be represented as a fraction. Rationals are automatically reduced to their simplest form.

#include <nira/rational.hpp>
...

nira::Rational value(10, 3); // 10 / 3
value += nira::Rational(1, 2); // Add one half, new value is 23 / 6
double real_value = value.real(); // Convert to floating-point

By default Rational uses int for the underlying integer type. You may supply any signed integer type as a template parameter:

nira::Rational<std::int64_t> big_value(389'674'130'590'529, 30'943'876'602'937);
nira::Rational<char> small_value(17, 13);

nira::FixedPoint

FixedPoint models a fixed point number, a number with a fixed number of fractional digits.

#include <nira/fixed_point.hpp>
...

nira::FixedPoint<2> value(10, 50); // 10.50
value += nira::FixedPoint<2>(1, 25); // Add 1.25, new value is 11.75
auto whole = value.whole(); // Get whole part
auto fractional = value.fractional(); // Get fractional part

By default FixedPoint uses int for the underlying integer type. You may supply any signed integer type as a template parameter:

nira::FixedPoint<4, std::int64_t> precise_value(123'456, 7890); // 123456.7890
nira::FixedPoint<1, char> small_value(12, 3); // 12.3