用嵌套环和一些numpy函数调用进行多处理

Question

我已经阅读了一些有关多处理的编码示例，并且对此很困惑。这是我人为的例子：

import numpy as np
def data_processing(x,y,z): return np.array([x,y])*(z**0.5)
def foo(n1,n2):
    final_result = {}
    for i in range(n1):
        result = np.zeros([n2,n2])
        for j1 in range(n2):
            for j2 in range(j1):
                temp= data_processing(j1,j2,i)
                result[j1,j2] = np.prod(temp)
        final_result[str(i)] = result
    return final_result
if __name__ == '__main__':
    X = foo(9,9)

如果我想在利用所有CPU内核时运行此代码，我应该更改？

Answer 1

maby这可以有所帮助。

import multiprocessing
import numpy as np
import time
import multiprocessing

def data_processing(x, y, z): return np.array([x, y]) * (z ** 0.5)

def foo(n1, n2, id=0, return_dict=[None]):
    final_result = {}
    for i in range(n1):
        result = np.zeros([n2, n2])
        for j1 in range(n2):
            for j2 in range(j1):
                temp = data_processing(j1, j2, i)
                result[j1, j2] = np.prod(temp)
        final_result[str(i)] = result
    return_dict[id] = final_result

stamp = time.time()
def pint(num):
    print(f'*Test [{num}] - seconds: {time.time() - stamp}')
for i in range(10):
    foo(90, 90)
pint(0)
stamp = time.time()
manager = multiprocessing.Manager()
return_dict = manager.dict()
processes = []
for i in range(10):
    p = multiprocessing.Process(target=foo, args=(90, 90, i, return_dict))
    processes.append(p)
    p.start()
for p in processes:
    p.join()
x0, x1, x2, x3, x4, x5, x6, x7, x8, x9 = return_dict.values()
pint(1)

我的输出是：

*Test [0] - seconds: 26.120166301727295
*Test [1] - seconds: 8.343111753463745
Process finished with exit code 0