v0.3.1

Added

• Complete NEON vectorization implementation for partition operations.
• Enhanced vectorized merge operation with optimized fast paths for homogeneous data segments.
• Specialized vector processing for different data patterns.

Improved

• Partition function now processes 4 elements at once using NEON SIMD.
• Merge function utilizes vectorized comparisons and bulk transfers.
• Better detection of sequential data patterns for faster processing.
• Significant performance boost especially for large datasets and reverse-sorted data.
• Updated benchmark results demonstrating up to 4.4x speedup for reverse-sorted data and up to 5.94x faster sorting than std::qsort.