python - 修复多处理进程超时时间

Question

我试图使用多处理模块设置并行作业的代码。我想在 10 秒内终止所有子作业。我在进程连接函数中使用超时参数编写了以下代码。但是，我发现所有作业的终止时间都取决于超时值和作业数量。如何计算准确的时间？

import time
from multiprocessing import Manager, Process

def f(x, dic):
    time.sleep(60)
    print "Time: ", x
    dic[x] = "Done"

mng = Manager()
dic = mng.dict()
jobs = [Process(target=f, args=(i, dic)) for i in range(50)]

for job in jobs:
    job.start()

for job in jobs:
    job.join(timeout=10)

for job in jobs:
    job.terminate()

for job in jobs:
    job.join()

print dic.keys()

Answer 1

您可以先计算十秒超时，然后在单个循环中等待并终止作业。仅当超时尚未到来时才等待。

end_at = time.time() + 10
while jobs:
    job = jobs.pop()
    delta = end_at - time.time()
    if delta > 0:
        job.join(timeout=delta)
    job.terminate()
    job.join()