Module simd 

SIMD-optimized vector operations

Provides hardware-accelerated distance calculations:

• AVX2/FMA on x86_64
• SSE on x86
• NEON on ARM64/Apple Silicon

Functions

dot_product_simd

SIMD-optimized dot product for cosine similarity Vectors should be normalized for cosine similarity Uses optimized scalar code with better pipelining (like Redis fallback)

l2_distance_simd

SIMD-optimized L2 distance (Euclidean)

norm_simd

SIMD-optimized vector norm (length)

norm_squared_simd

SIMD-optimized vector norm (squared length)