Module simd 

SIMD acceleration layer with runtime feature detection SIMD acceleration layer with runtime feature detection

Provides optimized vector and matrix operations using platform-specific SIMD instructions (AVX2/AVX-512 on x86_64, NEON/SVE on aarch64) with automatic scalar fallback for unsupported architectures.

Re-exports

pub use detection::SimdCapabilities;

Modules

detection

Runtime SIMD feature detection

matrix_ops

SIMD-accelerated matrix operations

vector_ops

SIMD-accelerated vector operations for CPU fallback paths