Parallel Processing

You should be able to treat models like any other function in terms of parallelization. This example shows using the built-in multiprocessing to do process-based parallelization of an explicit system.

from multiprocessing import Pool

import condor


class Model(condor.ExplicitSystem):
    x = input()
    output.y = -x**2 + 2*x + 1


with Pool(5) as p:
    models = p.map(Model, [1, 2, 3])

for model in models:
    print(model.input, model.output)

ModelInput(x=1) ModelOutput(y=array([2.]))
ModelInput(x=2) ModelOutput(y=array([1.]))
ModelInput(x=3) ModelOutput(y=array([-2.]))