vsum_f32

→float

float vsum_f32(const float * in, int n)

Single-precision array sum reduction.

Computes $\sum_{i=0}^{n-1} \text{in}_i$ via vDSP_sve (single-pass, pairwise tree-style accumulation — numerically more stable than a serial loop).

References

Accelerate.framework vDSP_sve.

inconst float*

Input buffer.

nstd::size_t

Element count.

float

Sum of all $n$ input elements.

Loading…