v0.2.0

Added

• Enhanced NEON vectorization that processes 4 integers simultaneously
• Adaptive algorithm selection that detects data patterns for optimal sorting
• Custom thread management based on array size and Apple Silicon core characteristics
• Specialized fast-path optimizations for common data patterns
• Benchmark suite for comparison against standard sorting algorithms
• Documentation for performance characteristics and optimization techniques

Improved

• Performance on large arrays (up to 30% faster for 1M+ elements)
• Memory efficiency through optimized cache line alignment
• Handling of nearly-sorted and reverse-sorted data, now up to 3.5x faster
• Work distribution across Apple Silicon's heterogeneous cores
• Branch prediction by optimizing decision paths in sorting algorithms
• Compiler optimizations specific to Apple Silicon

Fixed

• Memory access patterns to reduce cache misses on Apple Silicon
• Thread allocation issues on heavily loaded systems
• Performance degradation on certain data distribution patterns
• Inconsistent behavior with large arrays (>2M elements)

Performance

• Random data: Outperforms quicksort by up to 12%
• Nearly-sorted data: Up to 40% performance improvement
• Reverse-sorted data: Up to 3.5x faster processing
• Large arrays: Memory overhead reduced by 15% compared to previous version

Documentation

• Added detailed benchmarks and performance comparison tables
• Expanded explanation of Apple Silicon optimizations
• Improved usage examples and API documentation
• Added troubleshooting section for common issues