Performance of naive convolution against Python Numpy

[Eben60]

romainvieme, without actually checking your code, may I point you to a discussion which followed this post:

General questions from Python user

In the course of my weekend procrastinations I contrieved an example to check my statement: some home-made implementation of convolution in both languages. Make up some signal and kernel data in Python (unimportant part which is not benchmarked): import numpy as np def lorentz(x, gamma, mu=0): return (gamma**2/(((x-mu)**2)+gamma**2))/(np.pi*gamma) def signal(x): return lorentz(x, 0.1, -0.3) + lorentz(x, 0.03, 0.1) + lorentz(x, 0.2, 0.25) def gauss(x, sigma, mu=0.0): return np.e…