Python x+a*y vs daxpy

daxpy.py

import numpy as np
import perfplot
from scipy.linalg.blas import daxpy

a = 1.3


def np_axpy(data):
    x, y = data
    return a * x + y


def sp_daxpy(data):
    x, y = data
    return daxpy(x, y, a=a)


b = perfplot.bench(
    setup=lambda n: np.random.rand(2, n),
    kernels=[np_axpy, sp_daxpy],
    n_range=[2 ** k for k in range(24)],
)
b.save("out.png")
b.show()

See https://stackoverflow.com/a/67358748/353337 for a more comprehensive comparison.