v2.0.0

🚀 FFT Implementation v2.0.0 - Major Performance Update

🎯 Highlights

• 5x Performance Improvement with optimized algorithms
• Fixed Critical Linux Bug (GitHub Issue #1)
• GPU Acceleration support (CUDA + Metal)
• Automatic Algorithm Selection with new fft_auto() API

🐛 Critical Bug Fixes

• GitHub Issue #1: Fixed timer_t type conflict on Linux systems
  • Renamed custom timer type to fft_timer_t
  • Maintains cross-platform compatibility
• Radix-4 FFT: Enhanced reliability with 91.4% test pass rate
• Memory Management: Fixed leaks and improved error handling
• Numerical Stability: Improved accuracy across all algorithms

⚡ Performance Optimizations

• Twiddle Factors: Special case handling for common angles (0°, 90°, 180°, 270°)
• Bit Reversal: SIMD-friendly optimizations for sizes ≤ 256
• Compiler Hints: Branch prediction and force inlining for critical paths
• Memory Access: Cache-optimized patterns for better performance

🚀 New Features

• Automatic Selection: fft_auto() chooses optimal algorithm
• GPU Acceleration: CUDA and Metal Performance Shaders support
• Planning API: FFTW-inspired interface for repeated transforms
• Cross-Platform: Full Linux, macOS, and Windows support

📊 Quality Metrics

• Test Coverage: 92.7% pass rate (454/490 tests)
• Benchmarks: 83.3% pass rate with zero failures
• Numerical Accuracy: Errors in 1e-13 range for large transforms
• Memory Safety: Zero leaks with proper cleanup functions

📚 Documentation

• Complete Rewrite: ReadTheDocs integration
• Optimization Guide: New comprehensive performance guide
• API Reference: Updated with v2.0.0 features
• Migration Guide: Detailed upgrade instructions from v1.0.0

🔄 Breaking Changes

• API Redesign: Use fft_auto() or planning API instead of direct calls
• Direction Constants: Now use -1 for forward, +1 for inverse FFT
• Memory Allocation: Deprecated non-aligned allocation

📥 Installation

Quick Start

git clone https://github.com/muditbhargava66/FFT-implementation-in-C.git
cd FFT-implementation-in-C
make all

Using the New API

#include <fft_auto.h>

int main() {
    complex_t* signal = fft_alloc_complex(1024);
    // ... fill signal ...
    fft_auto(signal, signal, 1024, -1);  // Automatic algorithm selection
    fft_free(signal);
    return 0;
}

🙏 Acknowledgments

• Thanks to all contributors and testers
• Special thanks for reporting GitHub Issue #1
• Community feedback that drove v2.0.0 improvements

📖 Full Changelog

See CHANGELOG.md for complete details.

---

Full Changelog: v1.0.0...v2.0.0