Function vec_dot_s8 

pub fn vec_dot_s8(lhs: &[i8], rhs: &[i8], lhs_offset: i32) -> i32

Vectorized dot product for int8 slices with an input zero-point offset (lhs_offset).

Efficiently processes 4-element or 8-element chunks to trigger compiler auto-vectorization or ARM SIMD instructions (SMLAD on Cortex-M4/M7, vmladavaq on Cortex-M55/M85).