Faster KNN search in Manticore: 2-pass HNSW, batched distances, and AVX-512

Manticore has implemented significant performance optimizations for KNN (k-nearest neighbor) searches through advanced indexing techniques and hardware acceleration, enabling faster vector similarity searches critical for AI-powered applications like semantic search and recommendation engines. These improvements directly reduce query latency and infrastructure costs while supporting the growing demand for AI and machine learning workloads in enterprise applications. IT organizations leveraging vector databases should evaluate these enhancements to optimize their AI infrastructure performance and reduce computational overhead.