Freestanding, Portable C Math Library

A self-contained, strict IEEE-754 implementation of the standard C math library (fdlibm-based).

This library is designed for freestanding environments (Operating System kernels, embedded firmware, bootloaders) where the standard C library (libc / libm) is unavailable.

Features & Caveats

• Freestanding: Zero dependencies. Does not require <math.h> or standard headers.
• Strict IEEE-754: Implements proper handling for NaN, Inf, and signed zero.
• Portable C99: Written in pure C without assembly.
• Endianness: Uses uint64_t unions for bit-manipulation, making it Endian-Independent on all standard IEEE-754 architectures.
• Performance:
  • This is a software implementation. It is slightly slower (2x - 5x) than hardware-accelerated system libraries.
  • Use this for accuracy and portability, not for high-performance computing (HPC).

Building & Testing

This project uses CMake to ensure correct compilation across Windows (MSVC), Linux (GCC/Clang), and macOS.

Windows (PowerShell)

mkdir build
cd build

# Configure and Build Release (Optimized)
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build . --config Release

# Install the executable to the root folder
cmake --install . --config Release --prefix "dist"
Move-Item -Path "dist/bin/run_tests.exe" -Destination "./run_tests.exe" -Force
Remove-Item -Recurse -Force "dist"

# Run the tests
.\run_tests.exe

# (Optional) Cleanup:
cd ..
rm -r build

Linux / macOS (Bash)

mkdir -p build && cd build

# Configure and Build (Optimized)
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .

# Run the tests directly
./run_tests

# (Optional) Cleanup:
cd ..
rm -rf build

Usage

To use this in your own kernel or project, simply add math.c and math.h to your code.

// In your code
#include "math.h"

// code here...

double result = sin(1.57);