是否可以使用multiprocessing.Event为进程池实现同步屏障? [英] Is it possible to use multiprocessing.Event to implement a synchronization barrier for pool of processes?

问题描述

是否可以使用multiprocessing.Event为进程池实现同步屏障?我正在使用Python 2.7.但是，看起来事件未在进程之间共享. 我的代码有问题吗?

Is it possible to use multiprocessing.Event to implement a synchronization barrier for pool of processes? I am using Python 2.7. But it looks like the Event is not shared between processes. Is something wrong with my code?

def test_func(event):
    event.wait()
    return datetime.datetime.now()

def log_result(result):
    result_list.append(result)

if __name__ == '__main__':
    pool_size=10
    pool = multiprocessing.Pool(processes=pool_size)
    event = multiprocessing.Manager().Event()
    for _ in xrange(pool_size):
        pool.apply_async(test_func, args=(event,), callback=log_result)
    pool.close()
    pool.join()   
    time.sleep(5)
    event.set()

推荐答案

您有死锁:pool.join()等待启动的进程，所有进程都等待event.wait(). Sedd池加入:

You have dead-lock: pool.join() waits for started processes and all processes wait on event.wait(). Sedd pool join:

join()

Wait for the worker processes to exit

基本上，您必须将"pool.join()"移到"event.set()"之后，并且一切都应该没问题.

Basically you have to move "pool.join()" after "event.set()" and all should be OK.

这篇关于是否可以使用multiprocessing.Event为进程池实现同步屏障?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！