triple_accel

Delivers up to 20-30x faster edit distance calculations by leveraging AVX2 and SSE4.1 instructions, with automatic fallback to scalar code for unsupported CPUs, as highlighted in the README.

Supports Hamming, Levenshtein, and restricted Damerau-Levenshtein distances with customizable edit costs, enabling tailored string comparison for diverse use cases like bioinformatics.

Dynamically chooses optimal vector width and data structures based on input string lengths at runtime, ensuring maximum efficiency across short and long strings as described.

Lightweight with no heavy dependencies, making it portable and easy to compile even on machines without SIMD support, as noted in the features section.