【冰糖Python】python并行化-multiprocessing，joblib，numba

此处尝试multiprocessing，joblib，numba对实行for并行的作用

假设需要实现一个双重循环（可设置有返回值或无返回值）

for i in range(100000):
    t_list = 0
    for j in range(10000):
        t_list = i - j

1、不并行

为方便比较，定义了2个函数，采用了嵌套结构

def ForLoop():
    for i in range(100000):
       t = ForLoopSimple(i)
 
 
def ForLoopSimple(i):
    t_list = 0
    for j in range(10000):
        t_list = i - j
    return t_list
 
 
# origin (53 sec)
time_start = time.time()
ForLoop()
time_end = time.time()
print("---------------------")
print("origin time:", time_end-time_start)

---------------------
origin time: 56.11203336715698

2、multiprogressing

为multiprogressing单独定义一个脚本，采用5个线程，主要如下

def main():
    time_start = time.time()
 
    pool = multiprocessing.Pool(processes=5)
    result = pool.map(ForLoopSimple, rang