//! SIMD-accelerated batch operations over `&[f64]` slices.
//!
//! These functions complement the streaming `Next`-trait indicators with
//! vectorized paths for bulk array workloads (backtests, historical
//! analysis, vectorized recomputation). They use [`wide`] under the hood,
//! which compiles to AVX2 on modern x86_64, paired NEON on aarch64, and
//! a scalar fallback otherwise — correct on every target, fast where the
//! hardware supports it.
//!
//! Every SIMD function in this module has a `_scalar` reference
//! implementation used for parity testing. The two must agree to within
//! a small floating-point epsilon for all inputs; any divergence is a bug.
pub use ;
pub use ;
pub use ;
pub use ;