Function simd_sum_i64 

pub fn simd_sum_i64(data: &[i64]) -> i64

SIMD-style i64 sum using 4-wide accumulator lanes.

Uses wrapping_add to avoid panic on overflow (matches V8 BigInt semantics). For arrays up to ~2^62 elements of bounded values, no overflow occurs.

Performance

• Expected speedup: 2-4x over scalar for large arrays
• The 4-lane accumulator enables auto-vectorization on x86-64 and aarch64