Python 标准库 进程池介绍

有些情况下，所要完成的工作可以分解并独立地分布到多个工作进程，对于这种简单的情况，可以用Pool类来管理固定数目的工作进程。作业的返回值会收集并作为一个列表返回。（以下程序cpu数量为2，相关函数解释见python 进程池2 - Pool相关函数）。

 

1 import multiprocessing 2  3 def do_calculation(data): 4     return data*2 5 def start_process(): 6     print 'Starting',multiprocessing.current_process().name 7  8 if __name__=='__main__': 9     inputs=list(range(10))10     print 'Inputs  :',inputs11 12     builtin_output=map(do_calculation,inputs)13     print 'Build-In :', builtin_output14 15     pool_size=multiprocessing.cpu_count()*216     pool=multiprocessing.Pool(processes=pool_size,17         initializer=start_process,)18 19     pool_outputs=pool.map(do_calculation,inputs)20     pool.close()21     pool.join()22 23     print 'Pool  :',pool_outputs

 

其运行后的输出结果如下：

1 Inputs  : [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]2 Build-In : [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]3 Starting PoolWorker-24 Starting PoolWorker-15 Starting PoolWorker-36 Starting PoolWorker-47 Pool  : [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]

默认情况下，Pool会创建固定数目的工作进程，并向这些工作进程传递作业，直到再没有更多作业为止。maxtasksperchild参数为每个进程执行task的最大数目，设置maxtasksperchild参数可以告诉池在完成一定数量任务之后重新启动一个工作进程，来避免运行时间很长的工作进程消耗太多的系统资源。

maxtasksperchild is the number of tasks a worker process can complete before it will exit and be replaced with a fresh worker process, to enable unused resources to be freed. The default maxtasksperchild is None, which means worker processes will live as long as the pool.

Worker processes within a Pool typically live for the complete duration of the Pool’s work queue. A frequent pattern found in other systems (such as Apache, mod_wsgi, etc) to free resources held by workers is to allow a worker within a pool to complete only a set amount of work before being exiting, being cleaned up and a new process spawned to replace the old one. The maxtasksperchild argument to the Pool exposes this ability to the end user. 

notice：

python 2.6.6 multiprocessing.Pool没有maxtaskperchild参数，Pool(processes=None, initializer=None, initargs=())

python 2.7.3 Pool(processes=None, initializer=None, initargs=(), maxtasksperchild=None)

 1 import multiprocessing 2  3 def do_calculation(data): 4     return data*2 5 def start_process(): 6     print 'Starting',multiprocessing.current_process().name 7  8 if __name__=='__main__': 9     inputs=list(range(10))10     print 'Inputs  :',inputs11 12     builtin_output=map(do_calculation,inputs)13     print 'Build-In :', builtin_output14 15     pool_size=multiprocessing.cpu_count()*216     pool=multiprocessing.Pool(processes=pool_size,17         initializer=start_process,maxtasksperchild=2)18 19     pool_outputs=pool.map(do_calculation,inputs)20     pool.close()21     pool.join()22 23     print 'Pool  :',pool_outputs

其运行后的输出结果为：

 1 Inputs  : [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] 2 Build-In : [0, 2, 4, 6, 8, 10, 12, 14, 16, 18] 3 Starting PoolWorker-1 4 Starting PoolWorker-2 5 Starting PoolWorker-3 6 Starting PoolWorker-4 7 Starting PoolWorker-5 8 Starting PoolWorker-6 9 Starting PoolWorker-710 Starting PoolWorker-811 Pool  : [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]

池完成其所分配的任务时，即使没有更多的工作要做，也会重新启动工作进程。从这个输出可以看到，尽管只有10个任务，而且每个工作进程一次可以完成两个任务，但是这里创建了8个工作进程。

更多的时候，我们不仅需要多进程执行，还需要关注每个进程的执行结果。

 1 import multiprocessing 2 import time 3  4 def func(msg): 5     for i in xrange(3): 6         print msg 7         time.sleep(1) 8     return "done " + msg 9 10 if __name__ == "__main__":11     pool = multiprocessing.Pool(processes=4)12     result = []13     for i in xrange(10):14         msg = "hello %d" %(i)15         result.append(pool.apply_async(func, (msg, )))16     pool.close()17     pool.join()18     for res in result:19         print res.get()20     print "Sub-process(es) done."

 

参考：

《Python 标准库》 10.4.17 进程池(p445)

http://www.coder4.com/archives/3352

 

原文：http://www.cnblogs.com/congbo/archive/2012/08/23/2652433.html

来源：http://www.cnblogs.com/congbo/archive/2012/08/23/2652433.html

 

http://www.16kan.com/post/213223.html