Introduction

libfxd is a C++20 headers-only library that implements fixed-point types as a templated class. It's licensed under the Apache Public License 2.0.

Quick example

#include <iostream>
 
#include <fxd/fxd.hpp> // include everything
 
int main()
{
    using F = fxd::fixed<20, 12>;
 
    F a = 5;   // construct from integer
    F b = 2.5; // construct from floating point
    F c;       // uninitialized
 
    std::cout << "Enter a value: ";
    std::cin >> c;
 
    F d = (a + b + c) / 3;
 
    std::cout << "Average: " << d << std::endl;
    std::cout << "which has raw value: " << d.raw_value << std::endl;
 
    using F2 = fxd::fixed<10, 10>;
    auto x = a < F2{5};
    std::cout << x << std::endl;
}

Installation

Since it's a headers-only library, no compilation is necessary; the compiler just needs to locate the headers using an option like -I /path/to/libfxd/include.

To install the headers, use the standard Automake procedure:

./configure --prefix=/usr
make
sudo make install

See the INSTALL file and/or run ./configure --help for more details.

If the source code was checked out from the repository, you may need to first run the bootstrap script to create the configure script.