Function batch_sum_f64 

pub fn batch_sum_f64(values: &[f64]) -> f64

SIMD-friendly sum of f64 values.

Uses 4-way accumulator for instruction-level parallelism and auto-vectorization. For small slices (<4), falls back to direct sum.

Example

use presentar_core::simd::batch_sum_f64;
let values = vec![1.0, 2.0, 3.0, 4.0];
assert_eq!(batch_sum_f64(&values), 10.0);