Python中For循环多核并行优化问题：改写函数陷入无限循环

示例输入 #

C0x = [[2.969129, 3.065011, 3.294439],
       [3.087682, 1.674345, 2.930011],
       [3.355765, 2.863887, 2.588477]]

C0y = [[6.748940, 9.673919, 7.662236],
       [9.527495, 11.328892, 10.147356],
       [7.738452, 10.280770, 7.684719]]

C0xy = [[0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0]]

C0yx = [[0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0]]

Cxx1 = [[2.970383, 3.065395, 3.295505],
        [3.088111, 1.674282, 2.930299],
        [3.356823, 2.864131, 2.589346]]

Cyy1 = [[6.733786, 9.656677, 7.646381],
        [9.510379, 11.309615, 10.129516],
        [7.722598, 10.262803, 7.668157]]

Cxy1 = [[0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0]]

Cyx1 = [[0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0],
        [0.0, 0.0, 0.0]]

dCx = [array([[-0.88391347, -0.9165879 , -1.00445679],
        [-0.9102315 , -0.94390882, -1.03437531],
        [-0.99567547, -1.03249243, -1.13146452]]),
 array([[-0.93911752, -0.49819029, -0.88115751],
        [-0.51945798, -0.2755809 , -0.48739065],
        [-0.87413297, -0.46370905, -0.82018765]])]
dCy = [array([[4.54319981, 6.52256593, 5.20680836],
        [6.38565572, 9.16779592, 7.31840445],
        [5.19995576, 7.46547485, 5.95950073]]),
 array([[ 9.36071422, 11.09190807,  9.94111537],
        [10.98180048, 13.0128158 , 11.66270877],
        [ 9.93855597, 11.7766108 , 10.55478873]])]
dCxy = [array([[0., 0., 0.],
        [0., 0., 0.],
        [0., 0., 0.]]),
 array([[0., 0., 0.],
        [0., 0., 0.],
        [0., 0., 0.]])]
dCyx = [array([[0., 0., 0.],
        [0., 0., 0.],
        [0., 0., 0.]]),
 array([[0., 0., 0.],
        [0., 0., 0.],
        [0., 0., 0.]])]

原函数输出 #

A = [[26.203265, 17.026796],
     [17.026796, 11.763451],
     [1138.015961, 1913.702747],
     [1913.702747, 3238.589801],
     [0.0, 0.0],
     [0.0, 0.0],
     [0.0, 0.0],
     [0.0, 0.0]]

B = [-0.016327,     -0.011957,     -2.966823,     -5.028449,     0.0,     0.0,     0.0,     0.0]

问题描述 #

需通过multiprocessing模块加速适用于大输入的矩阵计算函数，已将原函数defineMatrices改写为多核版本defineMatrices_mp，但程序陷入无限循环，且未实现B矩阵的计算。要求最终得到与原函数完全一致的A、B矩阵结果。

内容的提问来源于stack exchange，提问作者Accelerator