multiprocessing
— Paralelismo baseado em processo¶
Código-fonte: Lib/multiprocessing/
Disponibilidade
Este módulo não tem suporte em plataformas móveis ou plataformas WebAssembly.
Introdução¶
multiprocessing
é um pacote que oferece suporte à invocação de processos utilizando uma API semelhante ao módulo threading
. O pacote multiprocessing
oferece simultaneamente concorrência local e remota, efetivamente contornando a trava global do interpretador, ao utilizar subprocessos ao invés de threads. Devido a isso, o módulo multiprocessing
permite ao programador aproveitar totalmente os múltiplos processadores de uma máquina. Ele funciona tanto em POSIX como em Windows.
O módulo multiprocessing
também introduz APIs que não têm análogos no módulo threading
. Um exemplo principal disso é o objeto Pool
que oferece um meio conveniente de paralelizar a execução de uma função em vários valores de entrada, distribuindo os dados de entrada entre processos (paralelismo de dados). O exemplo a seguir demonstra a prática comum de definir tais funções em um módulo para que os processos filhos possam importar esse módulo com sucesso. Este exemplo básico de paralelismo de dados usando Pool
,
from multiprocessing import Pool
def f(x):
return x*x
if __name__ == '__main__':
with Pool(5) as p:
print(p.map(f, [1, 2, 3]))
vai exibir na saída padrão
[1, 4, 9]
Ver também
concurrent.futures.ProcessPoolExecutor
oferece uma interface de nível mais alto para enviar tarefas para um processo em segundo plano sem bloquear a execução do processo de chamada. Comparado ao uso direto da interface Pool
, a API concurrent.futures
permite mais prontamente que o envio de trabalho para o pool de processos subjacente seja separado da espera pelos resultados.
A classe Process
¶
Em multiprocessing
, os processos são gerados criando um objeto Process
e então chamando seu método start()
. Process
segue a API de threading.Thread
. Um exemplo trivial de um programa multiprocesso é
from multiprocessing import Process
def f(name):
print('hello', name)
if __name__ == '__main__':
p = Process(target=f, args=('bob',))
p.start()
p.join()
Para mostrar os IDs de processo individuais envolvidos, aqui está um exemplo expandido:
from multiprocessing import Process
import os
def info(title):
print(title)
print('module name:', __name__)
print('parent process:', os.getppid())
print('process id:', os.getpid())
def f(name):
info('function f')
print('hello', name)
if __name__ == '__main__':
info('main line')
p = Process(target=f, args=('bob',))
p.start()
p.join()
Para uma explicação do porquê a parte if __name__ == '__main__'
é necessária, veja Programming guidelines.
Contextos e métodos de inicialização¶
Dependendo da plataforma, multiprocessing
suporta três maneiras de iniciar um processo. Estes métodos de início são
- spawn
O processo pai inicia um novo processo de interpretador Python. O processo filho herdará apenas os recursos necessários para executar o método
run()
do objeto do processo. Em particular, descritores de arquivo e identificadores desnecessários do processo pai não serão herdados. Iniciar um processo usando esse método é bem lento comparado a usar fork ou forkserver.Disponível em plataformas POSIX e Windows. O padrão no Windows e macOS.
- fork
O processo pai usa
os.fork()
para fazer um fork do interpretador Python. O processo filho, quando começa, é efetivamente idêntico ao processo pai. Todos os recursos do pai são herdados pelo processo filho. Observe que fazer um fork com segurança de um processo multithread é problemático.Disponível em sistemas POSIX. Atualmente o padrão em POSIX, exceto macOS.
Nota
O método de início padrão deixará de ser fork no Python 3.14. O código que requer fork deve especificar explicitamente isso via
get_context()
ouset_start_method()
.Alterado na versão 3.12: Se o Python for capaz de detectar que seu processo tem várias threads, a função
os.fork()
que esse método de início chama internamente levantaráDeprecationWarning
. Use um método de início diferente. Veja a documentação deos.fork()
para mais explicações.- forkserver
Quando o programa é inicializado e seleciona o método de início forkserver, um processo de servidor é gerado. A partir disso, sempre que um novo processo é necessário, o processo pai conecta-se ao servidor e solicita que um novo processo seja feito um fork. O processo fork do servidor é de thread única, a menos que bibliotecas do sistema ou importações pré-carregadas gerem threads como um efeito colateral; neste sentido, geralmente é seguro usar
os.fork()
. Nenhum recurso desnecessário é herdado.Disponível em plataformas POSIX que suportam a passagem de descritores de arquivo em Unix pipes, como o Linux.
Alterado na versão 3.4: spawn adicionado em todas as plataformas POSIX, e forkserver adicionado para algumas plataformas POSIX. Processos filhos não herdam mais todos os handles herdáveis dos pais no Windows.
Alterado na versão 3.8: No macOS, o método de início spawn agora é o padrão. O método de início fork deve ser considerado inseguro, pois pode levar a travamentos do subprocesso, pois as bibliotecas do sistema macOS podem iniciar threads. Veja bpo-33725.
No POSIX, usar os métodos de início spawn ou forkserver também iniciará um processo rastreador de recursos que rastreia os recursos de sistema nomeados não vinculados (como semáforos nomeados ou objetos SharedMemory
) criados por processos do programa. Quando todos os processos tiverem saído, o resource tracker desvincula qualquer objeto rastreado restante. Normalmente, não deve haver nenhum, mas se um processo foi morto por um sinal, pode haver alguns recursos “vazados”. (Nem os semáforos vazados nem os segmentos de memória compartilhada serão desvinculados automaticamente até a próxima reinicialização do sistema. Isso é problemático para ambos os objetos porque o sistema permite apenas um número limitado de semáforos nomeados, e os segmentos de memória compartilhada ocupam algum espaço na memória principal.)
Para selecionar um método de início, você usa set_start_method()
na cláusula if __name__ == '__main__'
do módulo principal. Por exemplo:
import multiprocessing as mp
def foo(q):
q.put('hello')
if __name__ == '__main__':
mp.set_start_method('spawn')
q = mp.Queue()
p = mp.Process(target=foo, args=(q,))
p.start()
print(q.get())
p.join()
set_start_method()
não deve ser usada mais de uma vez no programa.
Alternativamente, você pode usar get_context()
para obter um objeto de contexto. Objetos de contexto têm a mesma API que o módulo multiprocessing e permitem que se usem vários métodos de início no mesmo programa.
import multiprocessing as mp
def foo(q):
q.put('hello')
if __name__ == '__main__':
ctx = mp.get_context('spawn')
q = ctx.Queue()
p = ctx.Process(target=foo, args=(q,))
p.start()
print(q.get())
p.join()
Note que objetos relacionados a um contexto podem não ser compatíveis com processos para um contexto diferente. Em particular, travas criadas usando o contexto fork não podem ser passados para processos iniciados usando os métodos de início spawn ou forkserver.
Uma biblioteca que deseja utilizar um método de início específico provavelmente deve utilizar get_context()
para evitar interferir na escolha do usuário.
Aviso
Os métodos de início 'spawn'
e 'forkserver'
geralmente não podem ser usadas com executáveis “congelados” (por exemplo, binários produzidos por pacotes como PyInstaller e cx_Freeze) em sistemas POSIX. O método de início 'fork'
pode funcionar se o código não usar threads.
Trocando objetos entre processos¶
multiprocessing
tem suporte a dois tipos de canal de comunicação entre processos:
Filas
A classe
Queue
é quase um clone dequeue.Queue
. Por exemplo:from multiprocessing import Process, Queue def f(q): q.put([42, None, 'hello']) if __name__ == '__main__': q = Queue() p = Process(target=f, args=(q,)) p.start() print(q.get()) # prints "[42, None, 'hello']" p.join()As filas são seguras para threads e processos. Qualquer objeto colocado em uma fila
multiprocessing
será serializado.
Encadeamentos
A função
Pipe()
retorna um par de objetos de conexão conectados por um encadeamento que por padrão é duplex (bidirecional). Por exemplo:from multiprocessing import Process, Pipe def f(conn): conn.send([42, None, 'hello']) conn.close() if __name__ == '__main__': parent_conn, child_conn = Pipe() p = Process(target=f, args=(child_conn,)) p.start() print(parent_conn.recv()) # exibe "[42, None, 'hello']" p.join()Os dois objetos de conexão retornados por
Pipe()
representam as duas extremidades do encadeamento. Cada objeto de conexão tem os métodossend()
erecv()
(entre outros). Observe que os dados em um encadeamento podem ser corrompidos se dois processos (ou threads) tentarem ler ou gravar na mesma extremidade do encadeamento ao mesmo tempo. Claro que não há risco de corrupção de processos usando extremidades diferentes do encadeamento ao mesmo tempo.O método
send()
serializa o objeto erecv()
recria o objeto.
Sincronização entre processos¶
multiprocessing
contém equivalentes de todas as primitivas de sincronização de threading
. Por exemplo, pode-se usar uma trava para garantir que apenas um processo exiba na saída padrão por vez:
from multiprocessing import Process, Lock
def f(l, i):
l.acquire()
try:
print('hello world', i)
finally:
l.release()
if __name__ == '__main__':
lock = Lock()
for num in range(10):
Process(target=f, args=(lock, num)).start()
Sem utilizar a saída da trava dos diferentes processos, é possível que tudo fique confuso.
Usando um pool de workers¶
A classe Pool
representa um pool de processos de worker. Ela tem métodos que permitem que tarefas sejam descarregadas para os processos de worker de algumas maneiras diferentes.
Por exemplo:
from multiprocessing import Pool, TimeoutError
import time
import os
def f(x):
return x*x
if __name__ == '__main__':
# inicia 4 processos de trabalhador
with Pool(processes=4) as pool:
# exibe "[0, 1, 4,..., 81]"
print(pool.map(f, range(10)))
# exibe mesmo números em ordem arbitrária
for i in pool.imap_unordered(f, range(10)):
print(i)
# calcula "f(20)" assincronamente
res = pool.apply_async(f, (20,)) # executa em *apenas* um processo
print(res.get(timeout=1)) # exibe "400"
# calcula "os.getpid()" assincronamente
res = pool.apply_async(os.getpid, ()) # executa em *apenas* um procsso
print(res.get(timeout=1)) # exibe o PID daquele processo
# iniciando vários cálculos de forma assíncrona *pod* usar mais procssos
multiple_results = [pool.apply_async(os.getpid, ()) for i in range(4)]
print([res.get(timeout=1) for res in multiple_results])
# faz um único worker dormir por 10 segundos
res = pool.apply_async(time.sleep, (10,))
try:
print(res.get(timeout=1))
except TimeoutError:
print("We lacked patience and got a multiprocessing.TimeoutError")
print("For the moment, the pool remains available for more work")
# saindo o bloco 'with' parou o pool
print("Now the pool is closed and no longer available")
Observe que os métodos de um pool só devem ser usados pelo processo que o criou.
Nota
A funcionalidade dentro deste pacote requer que o módulo __main__
seja importável pelos filhos. Isso é abordado em Programming guidelines, mas vale a pena apontar aqui. Isso significa que alguns exemplos, como os exemplos multiprocessing.pool.Pool
não funcionarão no interpretador interativo. Por exemplo:
>>> from multiprocessing import Pool
>>> p = Pool(5)
>>> def f(x):
... return x*x
...
>>> with p:
... p.map(f, [1,2,3])
Process PoolWorker-1:
Process PoolWorker-2:
Process PoolWorker-3:
Traceback (most recent call last):
Traceback (most recent call last):
Traceback (most recent call last):
AttributeError: Can't get attribute 'f' on <module '__main__' (<class '_frozen_importlib.BuiltinImporter'>)>
AttributeError: Can't get attribute 'f' on <module '__main__' (<class '_frozen_importlib.BuiltinImporter'>)>
AttributeError: Can't get attribute 'f' on <module '__main__' (<class '_frozen_importlib.BuiltinImporter'>)>
(Se você tentar isso, na verdade, serão gerados três tracebaks completos intercalados de forma semi-aleatória, e então você pode ter que interromper o processo pai de alguma forma.)
Referência¶
O pacote multiprocessing
replica principalmente a API do módulo threading
.
Process
e exceções¶
- class multiprocessing.Process(group=None, target=None, name=None, args=(), kwargs={}, *, daemon=None)¶
Objetos processo representam atividades que são executadas em um processo separado. A classe
Process
possui equivalentes de todos os métodos dethreading.Thread
.O construtor deve sempre ser chamado com argumentos nomeados. group deve sempre ser
None
; ele existe somente para compatibilidade comthreading.Thread
. target é o objeto chamável a ser invocado pelo métodorun()
. O padrão éNone
, o que significa que nada é chamado. name é o nome do processo (vejaname
para mais detalhes). args é a tupla de argumento para a invocação de destino. kwargs é um dicionário de argumentos nomeados para a invocação de destino. Se fornecido, o argumento somente-nomeados daemon define o sinalizador do processodaemon
comoTrue
ouFalse
. SeNone
(o padrão), este sinalizador será herdado do processo de criação.Por padrão, nenhum argumento é passado para target. O argumento args, que tem como padrão
()
, pode ser usado para especificar uma lista ou tupla de argumentos a serem passados para target.Se uma subclasse substitui o construtor, ela deve certificar-se de invocar o construtor da classe base (
Process.__init__()
) antes de fazer qualquer outra coisa no processo.Alterado na versão 3.3: Adicionado o parâmetro daemon.
- run()¶
Método que representa a atividade do processo.
Você pode substituir esse método em uma subclasse. O método padrão
run()
invoca o objeto chamável passado ao construtor do objeto como o argumento alvo, se houver, com argumentos nomeados e sequenciais retirados dos argumentos args e kwargs, respectivamente.Usar uma lista ou tupla como argumento args passado para
Process
obtém o mesmo efeito.Exemplo:
>>> from multiprocessing import Process >>> p = Process(target=print, args=[1]) >>> p.run() 1 >>> p = Process(target=print, args=(1,)) >>> p.run() 1
- start()¶
Inicia a atividade do processo.
Isso deve ser chamado no máximo uma vez por objeto processo. Ele organiza para que o método
run()
do objeto seja invocado em um processo separado.
- join([timeout])¶
Se o argumento opcional timeout for
None
(o padrão), o método bloqueia até que o processo cujo métodojoin()
é chamado termine. Se timeout for um número positivo, ele bloqueia no máximo timeout segundos. Observe que o método retornaNone
se seu processo terminar ou se o método tiver tempo limite. Verifique oexitcode
do processo para determinar se ele terminou.Um processo pode ser usar “join” muitas vezes.
Um processo não pode se unir porque isso causaria um impasse. É um erro tentar se unir a um processo antes que ele tenha sido iniciado.
- name¶
O nome do processo. O nome é uma string usada apenas para fins de identificação. Não tem semântica. Vários processos podem receber o mesmo nome.
O nome inicial é definido pelo construtor. Se nenhum nome explícito for fornecido ao construtor, um nome do formato ‘Processo-N1:N2:…:Nk’ é construído, onde cada Nk é o N-ésimo filho de seu pai.
- is_alive()¶
Retorna se o processo está ativo.
Em termos gerais, um objeto processo está ativo desde o momento em que o método
start()
retorna até o término do processo filho.
- daemon¶
O sinalizador daemon do processo, um valor Booleano. Isso deve ser definido antes de
start()
ser chamado.O valor inicial é herdado do processo de criação.
Quando um processo sai, ele tenta encerrar todos os seus processos filhos daemônicos.
Note que um processo daemônico não tem permissão para criar processos filhos. Caso contrário, um processo daemônico deixaria seus filhos órfãos se ele fosse encerrado quando seu processo pai saísse. Além disso, esses não são daemons ou serviços Unix, eles são processos normais que serão encerrados (e em vez de usar “join”) se processos não daemônicos tiverem saído.
Além da API
threading.Thread
, os objetosProcess
também oferecem suporte aos seguintes atributos e métodos:- pid¶
Retorna o ID do processo. Antes do processo ser gerado, este será
None
.
- exitcode¶
O código de saída da criança. Este será
None
se o processo ainda não tiver terminado.Se o método
run()
da criança retornar normalmente, o código de saída será 0. Se ele terminar viasys.exit()
com um argumento inteiro N, o código de saída será N.Se a criança for encerrada devido a uma exceção não capturada em
run()
, o código de saída será 1. Se ela for encerrada pelo sinal N, o código de saída será o valor negativo -N.
- authkey¶
A chave de autenticação do processo (uma string de bytes).
Quando
multiprocessing
é inicializado, o processo principal recebe uma string aleatória usandoos.urandom()
.Quando um objeto
Process
é criado, ele herda a chave de autenticação do seu processo pai, embora isso possa ser alterado definindoauthkey
para outra sequência de bytes.Veja Authentication keys.
- sentinel¶
Um identificador numérico de um objeto do sistema que ficará “pronto” quando o processo terminar.
Você pode usar esse valor se quiser esperar por vários eventos ao mesmo tempo usando
multiprocessing.connection.wait()
. Caso contrário, chamarjoin()
é mais simples.No Windows, este é um identificador de sistema operacional utilizável com a família de chamadas de API
WaitForSingleObject
eWaitForMultipleObjects
. No POSIX, este é um descritor de arquivo utilizável com primitivos do móduloselect
.Adicionado na versão 3.3.
- terminate()¶
Termina o processo. No POSIX isso é feito usando o sinal
SIGTERM
; no WindowsTerminateProcess()
é usado. Note que os manipuladores de saída e cláusulas finally, etc., não serão executados.Observe que os processos descendentes do processo não serão encerrados — eles simplesmente ficarão órfãos.
Aviso
Se esse método for usado quando o processo associado estiver usando um encadeamento ou fila, então o encadeamento ou fila é passível de ser corrompido e pode se tornar inutilizável por outro processo. Similarmente, se o processo adquiriu um trava ou semáforo etc., então encerrá-lo é passível de causar impasse em outros processos.
- kill()¶
O mesmo que
terminate()
, mas usando o sinalSIGKILL
no POSIX.Adicionado na versão 3.7.
- close()¶
Fecha o objeto
Process
, liberando todos os recursos associados a ele.ValueError
é levantado se o processo subjacente ainda estiver em execução. Uma vez queclose()
retorne com sucesso, a maioria dos outros métodos e atributos do objetoProcess
levantaráValueError
.Adicionado na versão 3.7.
Observe que os métodos
start()
,join()
,is_alive()
,terminate()
eexitcode
devem ser chamados somente pelo processo que criou o objeto processo.Exemplo de uso de alguns dos métodos de
Process
:>>> import multiprocessing, time, signal >>> mp_context = multiprocessing.get_context('spawn') >>> p = mp_context.Process(target=time.sleep, args=(1000,)) >>> print(p, p.is_alive()) <...Process ... initial> False >>> p.start() >>> print(p, p.is_alive()) <...Process ... started> True >>> p.terminate() >>> time.sleep(0.1) >>> print(p, p.is_alive()) <...Process ... stopped exitcode=-SIGTERM> False >>> p.exitcode == -signal.SIGTERM True
- exception multiprocessing.ProcessError¶
A classe base de todas as exceções de
multiprocessing
.
- exception multiprocessing.BufferTooShort¶
Exceção levantada por
Connection.recv_bytes_into()
quando o objeto buffer fornecido é muito pequeno para a mensagem lida.Se
e
for uma instância deBufferTooShort
, entãoe.args[0]
retornará a mensagem como uma string de bytes.
- exception multiprocessing.AuthenticationError¶
Levantada quando há um erro de autenticação.
- exception multiprocessing.TimeoutError¶
Levantada por métodos com um tempo limite quando o tempo limite expira.
Encadeamentos e filas¶
Ao usar vários processos, geralmente é usada a passagem de mensagens para comunicação entre processos e evita-se ter que usar quaisquer primitivas de sincronização, como travas.
Para passar mensagens, pode-se usar Pipe()
(para uma conexão entre dois processos) ou uma fila (que permite múltiplos produtores e consumidores).
Os tipos Queue
, SimpleQueue
e JoinableQueue
são filas FIFO multiprodutoras e multiconsumidoras modeladas na classe queue.Queue
da biblioteca padrão. Elas diferem porque Queue
não tem os métodos task_done()
e join()
introduzidos na classe queue.Queue
do Python 2.5.
Se você usar JoinableQueue
, então você deve chamar JoinableQueue.task_done()
para cada tarefa removida da fila, caso contrário, o semáforo usado para contar o número de tarefas não concluídas pode eventualmente transbordar, levantando uma exceção.
Uma diferença de outras implementações de filas no Python é que as filas do multiprocessing
serializam todos os objetos que são colocados nelas usando pickle
. O objeto retornado pelo método get é um objeto recriado que não compartilha memória com o objeto original.
Observe que também é possível criar uma fila compartilhada usando um objeto gerenciador — veja Gerenciadores.
Nota
multiprocessing
usa as exceções usuais queue.Empty
e queue.Full
para sinalizar um tempo limite. Elas não estão disponíveis no espaço de nomes do multiprocessing
, então você precisa importá-las de queue
.
Nota
Quando um objeto é colocado em uma fila, o objeto é serializado com pickle e uma thread em segundo plano depois descarrega os dados serializados com pickle para um encadeamento subjacente. Isso tem algumas consequências que são um pouco surpreendentes, mas não devem causar nenhuma dificuldade prática – se elas realmente o incomodam, então você pode usar uma fila criada com um gerenciador.
Depois de colocar um objeto em uma fila vazia, pode haver um atraso infinitesimal antes que o método
empty()
da fila retorneFalse
e :meth:~Queue.get_nowait possa retornar sem levantarqueue.Empty
.Se vários processos estiverem enfileirando objetos, é possível que os objetos sejam recebidos na outra extremidade fora de ordem. No entanto, objetos enfileirados pelo mesmo processo sempre estarão na ordem esperada em relação uns aos outros.
Aviso
Se um processo for morto usando Process.terminate()
ou os.kill()
enquanto estiver tentando usar uma Queue
, os dados na fila provavelmente serão corrompidos. Isso pode fazer com que qualquer outro processo obtenha uma exceção quando tentar usar a fila mais tarde.
Aviso
Conforme mencionado acima, se um processo filho tiver colocado itens em uma fila (e não tiver usado JoinableQueue.cancel_join_thread
), esse processo não será encerrado até que todos os itens armazenados em buffer tenham sido liberados para o encadeamento.
This means that if you try joining that process you may get a deadlock unless you are sure that all items which have been put on the queue have been consumed. Similarly, if the child process is non-daemonic then the parent process may hang on exit when it tries to join all its non-daemonic children.
Note that a queue created using a manager does not have this issue. See Programming guidelines.
For an example of the usage of queues for interprocess communication see Exemplos.
- multiprocessing.Pipe([duplex])¶
Returns a pair
(conn1, conn2)
ofConnection
objects representing the ends of a pipe.If duplex is
True
(the default) then the pipe is bidirectional. If duplex isFalse
then the pipe is unidirectional:conn1
can only be used for receiving messages andconn2
can only be used for sending messages.The
send()
method serializes the the object usingpickle
and therecv()
re-creates the object.
- class multiprocessing.Queue([maxsize])¶
Returns a process shared queue implemented using a pipe and a few locks/semaphores. When a process first puts an item on the queue a feeder thread is started which transfers objects from a buffer into the pipe.
The usual
queue.Empty
andqueue.Full
exceptions from the standard library’squeue
module are raised to signal timeouts.Queue
implements all the methods ofqueue.Queue
except fortask_done()
andjoin()
.- qsize()¶
Return the approximate size of the queue. Because of multithreading/multiprocessing semantics, this number is not reliable.
Note that this may raise
NotImplementedError
on platforms like macOS wheresem_getvalue()
is not implemented.
- empty()¶
Return
True
if the queue is empty,False
otherwise. Because of multithreading/multiprocessing semantics, this is not reliable.May raise an
OSError
on closed queues. (not guaranteed)
- full()¶
Return
True
if the queue is full,False
otherwise. Because of multithreading/multiprocessing semantics, this is not reliable.
- put(obj[, block[, timeout]])¶
Put obj into the queue. If the optional argument block is
True
(the default) and timeout isNone
(the default), block if necessary until a free slot is available. If timeout is a positive number, it blocks at most timeout seconds and raises thequeue.Full
exception if no free slot was available within that time. Otherwise (block isFalse
), put an item on the queue if a free slot is immediately available, else raise thequeue.Full
exception (timeout is ignored in that case).Alterado na versão 3.8: If the queue is closed,
ValueError
is raised instead ofAssertionError
.
- put_nowait(obj)¶
Equivalent to
put(obj, False)
.
- get([block[, timeout]])¶
Remove and return an item from the queue. If optional args block is
True
(the default) and timeout isNone
(the default), block if necessary until an item is available. If timeout is a positive number, it blocks at most timeout seconds and raises thequeue.Empty
exception if no item was available within that time. Otherwise (block isFalse
), return an item if one is immediately available, else raise thequeue.Empty
exception (timeout is ignored in that case).Alterado na versão 3.8: If the queue is closed,
ValueError
is raised instead ofOSError
.
- get_nowait()¶
Equivalente a
get(False)
.
multiprocessing.Queue
has a few additional methods not found inqueue.Queue
. These methods are usually unnecessary for most code:- close()¶
Indicate that no more data will be put on this queue by the current process. The background thread will quit once it has flushed all buffered data to the pipe. This is called automatically when the queue is garbage collected.
- join_thread()¶
Join the background thread. This can only be used after
close()
has been called. It blocks until the background thread exits, ensuring that all data in the buffer has been flushed to the pipe.By default if a process is not the creator of the queue then on exit it will attempt to join the queue’s background thread. The process can call
cancel_join_thread()
to makejoin_thread()
do nothing.
- cancel_join_thread()¶
Prevent
join_thread()
from blocking. In particular, this prevents the background thread from being joined automatically when the process exits – seejoin_thread()
.A better name for this method might be
allow_exit_without_flush()
. It is likely to cause enqueued data to be lost, and you almost certainly will not need to use it. It is really only there if you need the current process to exit immediately without waiting to flush enqueued data to the underlying pipe, and you don’t care about lost data.
Nota
This class’s functionality requires a functioning shared semaphore implementation on the host operating system. Without one, the functionality in this class will be disabled, and attempts to instantiate a
Queue
will result in anImportError
. See bpo-3770 for additional information. The same holds true for any of the specialized queue types listed below.
- class multiprocessing.SimpleQueue¶
It is a simplified
Queue
type, very close to a lockedPipe
.- close()¶
Close the queue: release internal resources.
A queue must not be used anymore after it is closed. For example,
get()
,put()
andempty()
methods must no longer be called.Adicionado na versão 3.9.
- empty()¶
Retorna
True
se a fila estiver vazia,False
caso contrário.Always raises an
OSError
if the SimpleQueue is closed.
- get()¶
Remove and return an item from the queue.
- put(item)¶
Put item into the queue.
- class multiprocessing.JoinableQueue([maxsize])¶
JoinableQueue
, aQueue
subclass, is a queue which additionally hastask_done()
andjoin()
methods.- task_done()¶
Indicate that a formerly enqueued task is complete. Used by queue consumers. For each
get()
used to fetch a task, a subsequent call totask_done()
tells the queue that the processing on the task is complete.If a
join()
is currently blocking, it will resume when all items have been processed (meaning that atask_done()
call was received for every item that had beenput()
into the queue).Raises a
ValueError
if called more times than there were items placed in the queue.
- join()¶
Block until all items in the queue have been gotten and processed.
The count of unfinished tasks goes up whenever an item is added to the queue. The count goes down whenever a consumer calls
task_done()
to indicate that the item was retrieved and all work on it is complete. When the count of unfinished tasks drops to zero,join()
unblocks.
Diversos¶
- multiprocessing.active_children()¶
Return list of all live children of the current process.
Calling this has the side effect of “joining” any processes which have already finished.
- multiprocessing.cpu_count()¶
Return the number of CPUs in the system.
This number is not equivalent to the number of CPUs the current process can use. The number of usable CPUs can be obtained with
os.process_cpu_count()
(orlen(os.sched_getaffinity(0))
).When the number of CPUs cannot be determined a
NotImplementedError
is raised.Ver também
Alterado na versão 3.13: The return value can also be overridden using the
-X cpu_count
flag orPYTHON_CPU_COUNT
as this is merely a wrapper around theos
cpu count APIs.
- multiprocessing.current_process()¶
Return the
Process
object corresponding to the current process.An analogue of
threading.current_thread()
.
- multiprocessing.parent_process()¶
Return the
Process
object corresponding to the parent process of thecurrent_process()
. For the main process,parent_process
will beNone
.Adicionado na versão 3.8.
- multiprocessing.freeze_support()¶
Add support for when a program which uses
multiprocessing
has been frozen to produce a Windows executable. (Has been tested with py2exe, PyInstaller and cx_Freeze.)One needs to call this function straight after the
if __name__ == '__main__'
line of the main module. For example:from multiprocessing import Process, freeze_support def f(): print('hello world!') if __name__ == '__main__': freeze_support() Process(target=f).start()
If the
freeze_support()
line is omitted then trying to run the frozen executable will raiseRuntimeError
.Calling
freeze_support()
has no effect when invoked on any operating system other than Windows. In addition, if the module is being run normally by the Python interpreter on Windows (the program has not been frozen), thenfreeze_support()
has no effect.
- multiprocessing.get_all_start_methods()¶
Returns a list of the supported start methods, the first of which is the default. The possible start methods are
'fork'
,'spawn'
and'forkserver'
. Not all platforms support all methods. See Contextos e métodos de inicialização.Adicionado na versão 3.4.
- multiprocessing.get_context(method=None)¶
Return a context object which has the same attributes as the
multiprocessing
module.If method is
None
then the default context is returned. Otherwise method should be'fork'
,'spawn'
,'forkserver'
.ValueError
is raised if the specified start method is not available. See Contextos e métodos de inicialização.Adicionado na versão 3.4.
- multiprocessing.get_start_method(allow_none=False)¶
Return the name of start method used for starting processes.
If the start method has not been fixed and allow_none is false, then the start method is fixed to the default and the name is returned. If the start method has not been fixed and allow_none is true then
None
is returned.The return value can be
'fork'
,'spawn'
,'forkserver'
orNone
. See Contextos e métodos de inicialização.Adicionado na versão 3.4.
Alterado na versão 3.8: On macOS, the spawn start method is now the default. The fork start method should be considered unsafe as it can lead to crashes of the subprocess. See bpo-33725.
- multiprocessing.set_executable(executable)¶
Set the path of the Python interpreter to use when starting a child process. (By default
sys.executable
is used). Embedders will probably need to do some thing likeset_executable(os.path.join(sys.exec_prefix, 'pythonw.exe'))
before they can create child processes.
Alterado na versão 3.4: Now supported on POSIX when the
'spawn'
start method is used.Alterado na versão 3.11: Aceita um objeto caminho ou similar.
- multiprocessing.set_forkserver_preload(module_names)¶
Set a list of module names for the forkserver main process to attempt to import so that their already imported state is inherited by forked processes. Any
ImportError
when doing so is silently ignored. This can be used as a performance enhancement to avoid repeated work in every process.For this to work, it must be called before the forkserver process has been launched (before creating a
Pool
or starting aProcess
).Only meaningful when using the
'forkserver'
start method. See Contextos e métodos de inicialização.Adicionado na versão 3.4.
- multiprocessing.set_start_method(method, force=False)¶
Set the method which should be used to start child processes. The method argument can be
'fork'
,'spawn'
or'forkserver'
. RaisesRuntimeError
if the start method has already been set and force is notTrue
. If method isNone
and force isTrue
then the start method is set toNone
. If method isNone
and force isFalse
then the context is set to the default context.Note that this should be called at most once, and it should be protected inside the
if __name__ == '__main__'
clause of the main module.See Contextos e métodos de inicialização.
Adicionado na versão 3.4.
Nota
multiprocessing
contains no analogues of
threading.active_count()
, threading.enumerate()
,
threading.settrace()
, threading.setprofile()
,
threading.Timer
, or threading.local
.
Connection Objects¶
Connection objects allow the sending and receiving of picklable objects or strings. They can be thought of as message oriented connected sockets.
Connection objects are usually created using
Pipe
– see also
Listeners and Clients.
- class multiprocessing.connection.Connection¶
- send(obj)¶
Send an object to the other end of the connection which should be read using
recv()
.The object must be picklable. Very large pickles (approximately 32 MiB+, though it depends on the OS) may raise a
ValueError
exception.
- recv()¶
Return an object sent from the other end of the connection using
send()
. Blocks until there is something to receive. RaisesEOFError
if there is nothing left to receive and the other end was closed.
- fileno()¶
Return the file descriptor or handle used by the connection.
- close()¶
Close the connection.
This is called automatically when the connection is garbage collected.
- poll([timeout])¶
Return whether there is any data available to be read.
If timeout is not specified then it will return immediately. If timeout is a number then this specifies the maximum time in seconds to block. If timeout is
None
then an infinite timeout is used.Note that multiple connection objects may be polled at once by using
multiprocessing.connection.wait()
.
- send_bytes(buffer[, offset[, size]])¶
Send byte data from a bytes-like object as a complete message.
If offset is given then data is read from that position in buffer. If size is given then that many bytes will be read from buffer. Very large buffers (approximately 32 MiB+, though it depends on the OS) may raise a
ValueError
exception
- recv_bytes([maxlength])¶
Return a complete message of byte data sent from the other end of the connection as a string. Blocks until there is something to receive. Raises
EOFError
if there is nothing left to receive and the other end has closed.If maxlength is specified and the message is longer than maxlength then
OSError
is raised and the connection will no longer be readable.
- recv_bytes_into(buffer[, offset])¶
Read into buffer a complete message of byte data sent from the other end of the connection and return the number of bytes in the message. Blocks until there is something to receive. Raises
EOFError
if there is nothing left to receive and the other end was closed.buffer must be a writable bytes-like object. If offset is given then the message will be written into the buffer from that position. Offset must be a non-negative integer less than the length of buffer (in bytes).
If the buffer is too short then a
BufferTooShort
exception is raised and the complete message is available ase.args[0]
wheree
is the exception instance.
Alterado na versão 3.3: Connection objects themselves can now be transferred between processes using
Connection.send()
andConnection.recv()
.Connection objects also now support the context management protocol – see Tipos de Gerenciador de Contexto.
__enter__()
returns the connection object, and__exit__()
callsclose()
.
Por exemplo:
>>> from multiprocessing import Pipe
>>> a, b = Pipe()
>>> a.send([1, 'hello', None])
>>> b.recv()
[1, 'hello', None]
>>> b.send_bytes(b'thank you')
>>> a.recv_bytes()
b'thank you'
>>> import array
>>> arr1 = array.array('i', range(5))
>>> arr2 = array.array('i', [0] * 10)
>>> a.send_bytes(arr1)
>>> count = b.recv_bytes_into(arr2)
>>> assert count == len(arr1) * arr1.itemsize
>>> arr2
array('i', [0, 1, 2, 3, 4, 0, 0, 0, 0, 0])
Aviso
The Connection.recv()
method automatically unpickles the data it
receives, which can be a security risk unless you can trust the process
which sent the message.
Therefore, unless the connection object was produced using Pipe()
you
should only use the recv()
and send()
methods after performing some sort of authentication. See
Authentication keys.
Aviso
If a process is killed while it is trying to read or write to a pipe then the data in the pipe is likely to become corrupted, because it may become impossible to be sure where the message boundaries lie.
Synchronization primitives¶
Generally synchronization primitives are not as necessary in a multiprocess
program as they are in a multithreaded program. See the documentation for
threading
module.
Note that one can also create synchronization primitives by using a manager object – see Gerenciadores.
- class multiprocessing.Barrier(parties[, action[, timeout]])¶
A barrier object: a clone of
threading.Barrier
.Adicionado na versão 3.3.
- class multiprocessing.BoundedSemaphore([value])¶
A bounded semaphore object: a close analog of
threading.BoundedSemaphore
.A solitary difference from its close analog exists: its
acquire
method’s first argument is named block, as is consistent withLock.acquire()
.Nota
On macOS, this is indistinguishable from
Semaphore
becausesem_getvalue()
is not implemented on that platform.
- class multiprocessing.Condition([lock])¶
A condition variable: an alias for
threading.Condition
.If lock is specified then it should be a
Lock
orRLock
object frommultiprocessing
.Alterado na versão 3.3: The
wait_for()
method was added.
- class multiprocessing.Event¶
A clone of
threading.Event
.
- class multiprocessing.Lock¶
A non-recursive lock object: a close analog of
threading.Lock
. Once a process or thread has acquired a lock, subsequent attempts to acquire it from any process or thread will block until it is released; any process or thread may release it. The concepts and behaviors ofthreading.Lock
as it applies to threads are replicated here inmultiprocessing.Lock
as it applies to either processes or threads, except as noted.Note that
Lock
is actually a factory function which returns an instance ofmultiprocessing.synchronize.Lock
initialized with a default context.Lock
supports the context manager protocol and thus may be used inwith
statements.- acquire(block=True, timeout=None)¶
Acquire a lock, blocking or non-blocking.
With the block argument set to
True
(the default), the method call will block until the lock is in an unlocked state, then set it to locked and returnTrue
. Note that the name of this first argument differs from that inthreading.Lock.acquire()
.With the block argument set to
False
, the method call does not block. If the lock is currently in a locked state, returnFalse
; otherwise set the lock to a locked state and returnTrue
.When invoked with a positive, floating-point value for timeout, block for at most the number of seconds specified by timeout as long as the lock can not be acquired. Invocations with a negative value for timeout are equivalent to a timeout of zero. Invocations with a timeout value of
None
(the default) set the timeout period to infinite. Note that the treatment of negative orNone
values for timeout differs from the implemented behavior inthreading.Lock.acquire()
. The timeout argument has no practical implications if the block argument is set toFalse
and is thus ignored. ReturnsTrue
if the lock has been acquired orFalse
if the timeout period has elapsed.
- release()¶
Release a lock. This can be called from any process or thread, not only the process or thread which originally acquired the lock.
Behavior is the same as in
threading.Lock.release()
except that when invoked on an unlocked lock, aValueError
is raised.
- class multiprocessing.RLock¶
A recursive lock object: a close analog of
threading.RLock
. A recursive lock must be released by the process or thread that acquired it. Once a process or thread has acquired a recursive lock, the same process or thread may acquire it again without blocking; that process or thread must release it once for each time it has been acquired.Note that
RLock
is actually a factory function which returns an instance ofmultiprocessing.synchronize.RLock
initialized with a default context.RLock
supports the context manager protocol and thus may be used inwith
statements.- acquire(block=True, timeout=None)¶
Acquire a lock, blocking or non-blocking.
When invoked with the block argument set to
True
, block until the lock is in an unlocked state (not owned by any process or thread) unless the lock is already owned by the current process or thread. The current process or thread then takes ownership of the lock (if it does not already have ownership) and the recursion level inside the lock increments by one, resulting in a return value ofTrue
. Note that there are several differences in this first argument’s behavior compared to the implementation ofthreading.RLock.acquire()
, starting with the name of the argument itself.When invoked with the block argument set to
False
, do not block. If the lock has already been acquired (and thus is owned) by another process or thread, the current process or thread does not take ownership and the recursion level within the lock is not changed, resulting in a return value ofFalse
. If the lock is in an unlocked state, the current process or thread takes ownership and the recursion level is incremented, resulting in a return value ofTrue
.Use and behaviors of the timeout argument are the same as in
Lock.acquire()
. Note that some of these behaviors of timeout differ from the implemented behaviors inthreading.RLock.acquire()
.
- release()¶
Release a lock, decrementing the recursion level. If after the decrement the recursion level is zero, reset the lock to unlocked (not owned by any process or thread) and if any other processes or threads are blocked waiting for the lock to become unlocked, allow exactly one of them to proceed. If after the decrement the recursion level is still nonzero, the lock remains locked and owned by the calling process or thread.
Only call this method when the calling process or thread owns the lock. An
AssertionError
is raised if this method is called by a process or thread other than the owner or if the lock is in an unlocked (unowned) state. Note that the type of exception raised in this situation differs from the implemented behavior inthreading.RLock.release()
.
- class multiprocessing.Semaphore([value])¶
A semaphore object: a close analog of
threading.Semaphore
.A solitary difference from its close analog exists: its
acquire
method’s first argument is named block, as is consistent withLock.acquire()
.
Nota
On macOS, sem_timedwait
is unsupported, so calling acquire()
with
a timeout will emulate that function’s behavior using a sleeping loop.
Nota
Some of this package’s functionality requires a functioning shared semaphore
implementation on the host operating system. Without one, the
multiprocessing.synchronize
module will be disabled, and attempts to
import it will result in an ImportError
. See
bpo-3770 for additional information.
Gerenciadores¶
Managers provide a way to create data which can be shared between different processes, including sharing over a network between processes running on different machines. A manager object controls a server process which manages shared objects. Other processes can access the shared objects by using proxies.
- multiprocessing.Manager()¶
Returns a started
SyncManager
object which can be used for sharing objects between processes. The returned manager object corresponds to a spawned child process and has methods which will create shared objects and return corresponding proxies.
Manager processes will be shutdown as soon as they are garbage collected or
their parent process exits. The manager classes are defined in the
multiprocessing.managers
module:
- class multiprocessing.managers.BaseManager(address=None, authkey=None, serializer='pickle', ctx=None, *, shutdown_timeout=1.0)¶
Criando um objeto BaseManager.
Once created one should call
start()
orget_server().serve_forever()
to ensure that the manager object refers to a started manager process.address is the address on which the manager process listens for new connections. If address is
None
then an arbitrary one is chosen.authkey is the authentication key which will be used to check the validity of incoming connections to the server process. If authkey is
None
thencurrent_process().authkey
is used. Otherwise authkey is used and it must be a byte string.serializer must be
'pickle'
(usepickle
serialization) or'xmlrpclib'
(usexmlrpc.client
serialization).ctx is a context object, or
None
(use the current context). See theget_context()
function.shutdown_timeout is a timeout in seconds used to wait until the process used by the manager completes in the
shutdown()
method. If the shutdown times out, the process is terminated. If terminating the process also times out, the process is killed.Alterado na versão 3.11: Added the shutdown_timeout parameter.
- start([initializer[, initargs]])¶
Start a subprocess to start the manager. If initializer is not
None
then the subprocess will callinitializer(*initargs)
when it starts.
- get_server()¶
Returns a
Server
object which represents the actual server under the control of the Manager. TheServer
object supports theserve_forever()
method:>>> from multiprocessing.managers import BaseManager >>> manager = BaseManager(address=('', 50000), authkey=b'abc') >>> server = manager.get_server() >>> server.serve_forever()
Server
additionally has anaddress
attribute.
- connect()¶
Connect a local manager object to a remote manager process:
>>> from multiprocessing.managers import BaseManager >>> m = BaseManager(address=('127.0.0.1', 50000), authkey=b'abc') >>> m.connect()
- shutdown()¶
Stop the process used by the manager. This is only available if
start()
has been used to start the server process.This can be called multiple times.
- register(typeid[, callable[, proxytype[, exposed[, method_to_typeid[, create_method]]]]])¶
A classmethod which can be used for registering a type or callable with the manager class.
typeid is a “type identifier” which is used to identify a particular type of shared object. This must be a string.
callable is a callable used for creating objects for this type identifier. If a manager instance will be connected to the server using the
connect()
method, or if the create_method argument isFalse
then this can be left asNone
.proxytype is a subclass of
BaseProxy
which is used to create proxies for shared objects with this typeid. IfNone
then a proxy class is created automatically.exposed is used to specify a sequence of method names which proxies for this typeid should be allowed to access using
BaseProxy._callmethod()
. (If exposed isNone
thenproxytype._exposed_
is used instead if it exists.) In the case where no exposed list is specified, all “public methods” of the shared object will be accessible. (Here a “public method” means any attribute which has a__call__()
method and whose name does not begin with'_'
.)method_to_typeid is a mapping used to specify the return type of those exposed methods which should return a proxy. It maps method names to typeid strings. (If method_to_typeid is
None
thenproxytype._method_to_typeid_
is used instead if it exists.) If a method’s name is not a key of this mapping or if the mapping isNone
then the object returned by the method will be copied by value.create_method determines whether a method should be created with name typeid which can be used to tell the server process to create a new shared object and return a proxy for it. By default it is
True
.
BaseManager
instances also have one read-only property:- address¶
The address used by the manager.
Alterado na versão 3.3: Manager objects support the context management protocol – see Tipos de Gerenciador de Contexto.
__enter__()
starts the server process (if it has not already started) and then returns the manager object.__exit__()
callsshutdown()
.In previous versions
__enter__()
did not start the manager’s server process if it was not already started.
- class multiprocessing.managers.SyncManager¶
A subclass of
BaseManager
which can be used for the synchronization of processes. Objects of this type are returned bymultiprocessing.Manager()
.Its methods create and return Proxy Objects for a number of commonly used data types to be synchronized across processes. This notably includes shared lists and dictionaries.
- Barrier(parties[, action[, timeout]])¶
Create a shared
threading.Barrier
object and return a proxy for it.Adicionado na versão 3.3.
- BoundedSemaphore([value])¶
Create a shared
threading.BoundedSemaphore
object and return a proxy for it.
- Condition([lock])¶
Create a shared
threading.Condition
object and return a proxy for it.If lock is supplied then it should be a proxy for a
threading.Lock
orthreading.RLock
object.Alterado na versão 3.3: The
wait_for()
method was added.
- Event()¶
Create a shared
threading.Event
object and return a proxy for it.
- Lock()¶
Create a shared
threading.Lock
object and return a proxy for it.
- Queue([maxsize])¶
Create a shared
queue.Queue
object and return a proxy for it.
- RLock()¶
Create a shared
threading.RLock
object and return a proxy for it.
- Semaphore([value])¶
Create a shared
threading.Semaphore
object and return a proxy for it.
- Array(typecode, sequence)¶
Create an array and return a proxy for it.
- Value(typecode, value)¶
Create an object with a writable
value
attribute and return a proxy for it.
Alterado na versão 3.6: Shared objects are capable of being nested. For example, a shared container object such as a shared list can contain other shared objects which will all be managed and synchronized by the
SyncManager
.
- class multiprocessing.managers.Namespace¶
A type that can register with
SyncManager
.A namespace object has no public methods, but does have writable attributes. Its representation shows the values of its attributes.
However, when using a proxy for a namespace object, an attribute beginning with
'_'
will be an attribute of the proxy and not an attribute of the referent:>>> mp_context = multiprocessing.get_context('spawn') >>> manager = mp_context.Manager() >>> Global = manager.Namespace() >>> Global.x = 10 >>> Global.y = 'hello' >>> Global._z = 12.3 # this is an attribute of the proxy >>> print(Global) Namespace(x=10, y='hello')
Customized managers¶
To create one’s own manager, one creates a subclass of BaseManager
and
uses the register()
classmethod to register new types or
callables with the manager class. For example:
from multiprocessing.managers import BaseManager
class MathsClass:
def add(self, x, y):
return x + y
def mul(self, x, y):
return x * y
class MyManager(BaseManager):
pass
MyManager.register('Maths', MathsClass)
if __name__ == '__main__':
with MyManager() as manager:
maths = manager.Maths()
print(maths.add(4, 3)) # prints 7
print(maths.mul(7, 8)) # prints 56
Using a remote manager¶
It is possible to run a manager server on one machine and have clients use it from other machines (assuming that the firewalls involved allow it).
Running the following commands creates a server for a single shared queue which remote clients can access:
>>> from multiprocessing.managers import BaseManager
>>> from queue import Queue
>>> queue = Queue()
>>> class QueueManager(BaseManager): pass
>>> QueueManager.register('get_queue', callable=lambda:queue)
>>> m = QueueManager(address=('', 50000), authkey=b'abracadabra')
>>> s = m.get_server()
>>> s.serve_forever()
One client can access the server as follows:
>>> from multiprocessing.managers import BaseManager
>>> class QueueManager(BaseManager): pass
>>> QueueManager.register('get_queue')
>>> m = QueueManager(address=('foo.bar.org', 50000), authkey=b'abracadabra')
>>> m.connect()
>>> queue = m.get_queue()
>>> queue.put('hello')
Another client can also use it:
>>> from multiprocessing.managers import BaseManager
>>> class QueueManager(BaseManager): pass
>>> QueueManager.register('get_queue')
>>> m = QueueManager(address=('foo.bar.org', 50000), authkey=b'abracadabra')
>>> m.connect()
>>> queue = m.get_queue()
>>> queue.get()
'hello'
Local processes can also access that queue, using the code from above on the client to access it remotely:
>>> from multiprocessing import Process, Queue
>>> from multiprocessing.managers import BaseManager
>>> class Worker(Process):
... def __init__(self, q):
... self.q = q
... super().__init__()
... def run(self):
... self.q.put('local hello')
...
>>> queue = Queue()
>>> w = Worker(queue)
>>> w.start()
>>> class QueueManager(BaseManager): pass
...
>>> QueueManager.register('get_queue', callable=lambda: queue)
>>> m = QueueManager(address=('', 50000), authkey=b'abracadabra')
>>> s = m.get_server()
>>> s.serve_forever()
Proxy Objects¶
A proxy is an object which refers to a shared object which lives (presumably) in a different process. The shared object is said to be the referent of the proxy. Multiple proxy objects may have the same referent.
A proxy object has methods which invoke corresponding methods of its referent (although not every method of the referent will necessarily be available through the proxy). In this way, a proxy can be used just like its referent can:
>>> mp_context = multiprocessing.get_context('spawn')
>>> manager = mp_context.Manager()
>>> l = manager.list([i*i for i in range(10)])
>>> print(l)
[0, 1, 4, 9, 16, 25, 36, 49, 64, 81]
>>> print(repr(l))
<ListProxy object, typeid 'list' at 0x...>
>>> l[4]
16
>>> l[2:5]
[4, 9, 16]
Notice that applying str()
to a proxy will return the representation of
the referent, whereas applying repr()
will return the representation of
the proxy.
An important feature of proxy objects is that they are picklable so they can be passed between processes. As such, a referent can contain Proxy Objects. This permits nesting of these managed lists, dicts, and other Proxy Objects:
>>> a = manager.list()
>>> b = manager.list()
>>> a.append(b) # referent of a now contains referent of b
>>> print(a, b)
[<ListProxy object, typeid 'list' at ...>] []
>>> b.append('hello')
>>> print(a[0], b)
['hello'] ['hello']
Similarly, dict and list proxies may be nested inside one another:
>>> l_outer = manager.list([ manager.dict() for i in range(2) ])
>>> d_first_inner = l_outer[0]
>>> d_first_inner['a'] = 1
>>> d_first_inner['b'] = 2
>>> l_outer[1]['c'] = 3
>>> l_outer[1]['z'] = 26
>>> print(l_outer[0])
{'a': 1, 'b': 2}
>>> print(l_outer[1])
{'c': 3, 'z': 26}
If standard (non-proxy) list
or dict
objects are contained
in a referent, modifications to those mutable values will not be propagated
through the manager because the proxy has no way of knowing when the values
contained within are modified. However, storing a value in a container proxy
(which triggers a __setitem__
on the proxy object) does propagate through
the manager and so to effectively modify such an item, one could re-assign the
modified value to the container proxy:
# create a list proxy and append a mutable object (a dictionary)
lproxy = manager.list()
lproxy.append({})
# now mutate the dictionary
d = lproxy[0]
d['a'] = 1
d['b'] = 2
# at this point, the changes to d are not yet synced, but by
# updating the dictionary, the proxy is notified of the change
lproxy[0] = d
This approach is perhaps less convenient than employing nested Proxy Objects for most use cases but also demonstrates a level of control over the synchronization.
Nota
The proxy types in multiprocessing
do nothing to support comparisons
by value. So, for instance, we have:
>>> manager.list([1,2,3]) == [1,2,3]
False
One should just use a copy of the referent instead when making comparisons.
- class multiprocessing.managers.BaseProxy¶
Proxy objects are instances of subclasses of
BaseProxy
.- _callmethod(methodname[, args[, kwds]])¶
Call and return the result of a method of the proxy’s referent.
If
proxy
is a proxy whose referent isobj
then the expressionproxy._callmethod(methodname, args, kwds)
will evaluate the expression
getattr(obj, methodname)(*args, **kwds)
in the manager’s process.
The returned value will be a copy of the result of the call or a proxy to a new shared object – see documentation for the method_to_typeid argument of
BaseManager.register()
.If an exception is raised by the call, then is re-raised by
_callmethod()
. If some other exception is raised in the manager’s process then this is converted into aRemoteError
exception and is raised by_callmethod()
.Note in particular that an exception will be raised if methodname has not been exposed.
An example of the usage of
_callmethod()
:>>> l = manager.list(range(10)) >>> l._callmethod('__len__') 10 >>> l._callmethod('__getitem__', (slice(2, 7),)) # equivalent to l[2:7] [2, 3, 4, 5, 6] >>> l._callmethod('__getitem__', (20,)) # equivalent to l[20] Traceback (most recent call last): ... IndexError: list index out of range
- _getvalue()¶
Return a copy of the referent.
If the referent is unpicklable then this will raise an exception.
- __repr__()¶
Return a representation of the proxy object.
- __str__()¶
Return the representation of the referent.
Limpeza¶
A proxy object uses a weakref callback so that when it gets garbage collected it deregisters itself from the manager which owns its referent.
A shared object gets deleted from the manager process when there are no longer any proxies referring to it.
Process Pools¶
One can create a pool of processes which will carry out tasks submitted to it
with the Pool
class.
- class multiprocessing.pool.Pool([processes[, initializer[, initargs[, maxtasksperchild[, context]]]]])¶
A process pool object which controls a pool of worker processes to which jobs can be submitted. It supports asynchronous results with timeouts and callbacks and has a parallel map implementation.
processes is the number of worker processes to use. If processes is
None
then the number returned byos.process_cpu_count()
is used.If initializer is not
None
then each worker process will callinitializer(*initargs)
when it starts.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.context can be used to specify the context used for starting the worker processes. Usually a pool is created using the function
multiprocessing.Pool()
or thePool()
method of a context object. In both cases context is set appropriately.Note that the methods of the pool object should only be called by the process which created the pool.
Aviso
multiprocessing.pool
objects have internal resources that need to be properly managed (like any other resource) by using the pool as a context manager or by callingclose()
andterminate()
manually. Failure to do this can lead to the process hanging on finalization.Note that it is not correct to rely on the garbage collector to destroy the pool as CPython does not assure that the finalizer of the pool will be called (see
object.__del__()
for more information).Alterado na versão 3.2: Added the maxtasksperchild parameter.
Alterado na versão 3.4: Adicionado o parâmetro context.
Alterado na versão 3.13: processes uses
os.process_cpu_count()
by default, instead ofos.cpu_count()
.Nota
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 thePool
exposes this ability to the end user.- apply(func[, args[, kwds]])¶
Call func with arguments args and keyword arguments kwds. It blocks until the result is ready. Given this blocks,
apply_async()
is better suited for performing work in parallel. Additionally, func is only executed in one of the workers of the pool.
- apply_async(func[, args[, kwds[, callback[, error_callback]]]])¶
A variant of the
apply()
method which returns aAsyncResult
object.If callback is specified then it should be a callable which accepts a single argument. When the result becomes ready callback is applied to it, that is unless the call failed, in which case the error_callback is applied instead.
If error_callback is specified then it should be a callable which accepts a single argument. If the target function fails, then the error_callback is called with the exception instance.
Callbacks should complete immediately since otherwise the thread which handles the results will get blocked.
- map(func, iterable[, chunksize])¶
A parallel equivalent of the
map()
built-in function (it supports only one iterable argument though, for multiple iterables seestarmap()
). It blocks until the result is ready.This method chops the iterable into a number of chunks which it submits to the process pool as separate tasks. The (approximate) size of these chunks can be specified by setting chunksize to a positive integer.
Note that it may cause high memory usage for very long iterables. Consider using
imap()
orimap_unordered()
with explicit chunksize option for better efficiency.
- map_async(func, iterable[, chunksize[, callback[, error_callback]]])¶
A variant of the
map()
method which returns aAsyncResult
object.If callback is specified then it should be a callable which accepts a single argument. When the result becomes ready callback is applied to it, that is unless the call failed, in which case the error_callback is applied instead.
If error_callback is specified then it should be a callable which accepts a single argument. If the target function fails, then the error_callback is called with the exception instance.
Callbacks should complete immediately since otherwise the thread which handles the results will get blocked.
- imap(func, iterable[, chunksize])¶
A lazier version of
map()
.The chunksize argument is the same as the one used by the
map()
method. For very long iterables using a large value for chunksize can make the job complete much faster than using the default value of1
.Also if chunksize is
1
then thenext()
method of the iterator returned by theimap()
method has an optional timeout parameter:next(timeout)
will raisemultiprocessing.TimeoutError
if the result cannot be returned within timeout seconds.
- imap_unordered(func, iterable[, chunksize])¶
The same as
imap()
except that the ordering of the results from the returned iterator should be considered arbitrary. (Only when there is only one worker process is the order guaranteed to be “correct”.)
- starmap(func, iterable[, chunksize])¶
Like
map()
except that the elements of the iterable are expected to be iterables that are unpacked as arguments.Hence an iterable of
[(1,2), (3, 4)]
results in[func(1,2), func(3,4)]
.Adicionado na versão 3.3.
- starmap_async(func, iterable[, chunksize[, callback[, error_callback]]])¶
A combination of
starmap()
andmap_async()
that iterates over iterable of iterables and calls func with the iterables unpacked. Returns a result object.Adicionado na versão 3.3.
- close()¶
Prevents any more tasks from being submitted to the pool. Once all the tasks have been completed the worker processes will exit.
- terminate()¶
Stops the worker processes immediately without completing outstanding work. When the pool object is garbage collected
terminate()
will be called immediately.
- join()¶
Wait for the worker processes to exit. One must call
close()
orterminate()
before usingjoin()
.
Alterado na versão 3.3: Pool objects now support the context management protocol – see Tipos de Gerenciador de Contexto.
__enter__()
returns the pool object, and__exit__()
callsterminate()
.
- class multiprocessing.pool.AsyncResult¶
The class of the result returned by
Pool.apply_async()
andPool.map_async()
.- get([timeout])¶
Return the result when it arrives. If timeout is not
None
and the result does not arrive within timeout seconds thenmultiprocessing.TimeoutError
is raised. If the remote call raised an exception then that exception will be reraised byget()
.
- wait([timeout])¶
Wait until the result is available or until timeout seconds pass.
- ready()¶
Return whether the call has completed.
- successful()¶
Return whether the call completed without raising an exception. Will raise
ValueError
if the result is not ready.Alterado na versão 3.7: If the result is not ready,
ValueError
is raised instead ofAssertionError
.
The following example demonstrates the use of a pool:
from multiprocessing import Pool
import time
def f(x):
return x*x
if __name__ == '__main__':
with Pool(processes=4) as pool: # start 4 worker processes
result = pool.apply_async(f, (10,)) # evaluate "f(10)" asynchronously in a single process
print(result.get(timeout=1)) # prints "100" unless your computer is *very* slow
print(pool.map(f, range(10))) # prints "[0, 1, 4,..., 81]"
it = pool.imap(f, range(10))
print(next(it)) # prints "0"
print(next(it)) # prints "1"
print(it.next(timeout=1)) # prints "4" unless your computer is *very* slow
result = pool.apply_async(time.sleep, (10,))
print(result.get(timeout=1)) # raises multiprocessing.TimeoutError
Listeners and Clients¶
Usually message passing between processes is done using queues or by using
Connection
objects returned by
Pipe()
.
However, the multiprocessing.connection
module allows some extra
flexibility. It basically gives a high level message oriented API for dealing
with sockets or Windows named pipes. It also has support for digest
authentication using the hmac
module, and for polling
multiple connections at the same time.
- multiprocessing.connection.deliver_challenge(connection, authkey)¶
Send a randomly generated message to the other end of the connection and wait for a reply.
If the reply matches the digest of the message using authkey as the key then a welcome message is sent to the other end of the connection. Otherwise
AuthenticationError
is raised.
- multiprocessing.connection.answer_challenge(connection, authkey)¶
Receive a message, calculate the digest of the message using authkey as the key, and then send the digest back.
If a welcome message is not received, then
AuthenticationError
is raised.
- multiprocessing.connection.Client(address[, family[, authkey]])¶
Attempt to set up a connection to the listener which is using address address, returning a
Connection
.The type of the connection is determined by family argument, but this can generally be omitted since it can usually be inferred from the format of address. (See Formatos de Endereços)
If authkey is given and not
None
, it should be a byte string and will be used as the secret key for an HMAC-based authentication challenge. No authentication is done if authkey isNone
.AuthenticationError
is raised if authentication fails. See Authentication keys.
- class multiprocessing.connection.Listener([address[, family[, backlog[, authkey]]]])¶
A wrapper for a bound socket or Windows named pipe which is ‘listening’ for connections.
address is the address to be used by the bound socket or named pipe of the listener object.
Nota
If an address of ‘0.0.0.0’ is used, the address will not be a connectable end point on Windows. If you require a connectable end-point, you should use ‘127.0.0.1’.
family is the type of socket (or named pipe) to use. This can be one of the strings
'AF_INET'
(for a TCP socket),'AF_UNIX'
(for a Unix domain socket) or'AF_PIPE'
(for a Windows named pipe). Of these only the first is guaranteed to be available. If family isNone
then the family is inferred from the format of address. If address is alsoNone
then a default is chosen. This default is the family which is assumed to be the fastest available. See Formatos de Endereços. Note that if family is'AF_UNIX'
and address isNone
then the socket will be created in a private temporary directory created usingtempfile.mkstemp()
.If the listener object uses a socket then backlog (1 by default) is passed to the
listen()
method of the socket once it has been bound.If authkey is given and not
None
, it should be a byte string and will be used as the secret key for an HMAC-based authentication challenge. No authentication is done if authkey isNone
.AuthenticationError
is raised if authentication fails. See Authentication keys.- accept()¶
Accept a connection on the bound socket or named pipe of the listener object and return a
Connection
object. If authentication is attempted and fails, thenAuthenticationError
is raised.
- close()¶
Close the bound socket or named pipe of the listener object. This is called automatically when the listener is garbage collected. However it is advisable to call it explicitly.
Listener objects have the following read-only properties:
- address¶
The address which is being used by the Listener object.
- last_accepted¶
The address from which the last accepted connection came. If this is unavailable then it is
None
.
Alterado na versão 3.3: Listener objects now support the context management protocol – see Tipos de Gerenciador de Contexto.
__enter__()
returns the listener object, and__exit__()
callsclose()
.
- multiprocessing.connection.wait(object_list, timeout=None)¶
Wait till an object in object_list is ready. Returns the list of those objects in object_list which are ready. If timeout is a float then the call blocks for at most that many seconds. If timeout is
None
then it will block for an unlimited period. A negative timeout is equivalent to a zero timeout.For both POSIX and Windows, an object can appear in object_list if it is
a readable
Connection
object;a connected and readable
socket.socket
object; or
A connection or socket object is ready when there is data available to be read from it, or the other end has been closed.
POSIX:
wait(object_list, timeout)
almost equivalentselect.select(object_list, [], [], timeout)
. The difference is that, ifselect.select()
is interrupted by a signal, it can raiseOSError
with an error number ofEINTR
, whereaswait()
will not.Windows: An item in object_list must either be an integer handle which is waitable (according to the definition used by the documentation of the Win32 function
WaitForMultipleObjects()
) or it can be an object with afileno()
method which returns a socket handle or pipe handle. (Note that pipe handles and socket handles are not waitable handles.)Adicionado na versão 3.3.
Examples
The following server code creates a listener which uses 'secret password'
as
an authentication key. It then waits for a connection and sends some data to
the client:
from multiprocessing.connection import Listener
from array import array
address = ('localhost', 6000) # family is deduced to be 'AF_INET'
with Listener(address, authkey=b'secret password') as listener:
with listener.accept() as conn:
print('connection accepted from', listener.last_accepted)
conn.send([2.25, None, 'junk', float])
conn.send_bytes(b'hello')
conn.send_bytes(array('i', [42, 1729]))
The following code connects to the server and receives some data from the server:
from multiprocessing.connection import Client
from array import array
address = ('localhost', 6000)
with Client(address, authkey=b'secret password') as conn:
print(conn.recv()) # => [2.25, None, 'junk', float]
print(conn.recv_bytes()) # => 'hello'
arr = array('i', [0, 0, 0, 0, 0])
print(conn.recv_bytes_into(arr)) # => 8
print(arr) # => array('i', [42, 1729, 0, 0, 0])
The following code uses wait()
to
wait for messages from multiple processes at once:
from multiprocessing import Process, Pipe, current_process
from multiprocessing.connection import wait
def foo(w):
for i in range(10):
w.send((i, current_process().name))
w.close()
if __name__ == '__main__':
readers = []
for i in range(4):
r, w = Pipe(duplex=False)
readers.append(r)
p = Process(target=foo, args=(w,))
p.start()
# We close the writable end of the pipe now to be sure that
# p is the only process which owns a handle for it. This
# ensures that when p closes its handle for the writable end,
# wait() will promptly report the readable end as being ready.
w.close()
while readers:
for r in wait(readers):
try:
msg = r.recv()
except EOFError:
readers.remove(r)
else:
print(msg)
Formatos de Endereços¶
Um endereço
'AF_INET'
é uma tupla na forma de(hostname, port)
sendo hostname uma string e port um inteiro.An
'AF_UNIX'
address is a string representing a filename on the filesystem.An
'AF_PIPE'
address is a string of the formr'\\.\pipe\PipeName'
. To useClient()
to connect to a named pipe on a remote computer called ServerName one should use an address of the formr'\\ServerName\pipe\PipeName'
instead.
Note that any string beginning with two backslashes is assumed by default to be
an 'AF_PIPE'
address rather than an 'AF_UNIX'
address.
Authentication keys¶
When one uses Connection.recv
, the
data received is automatically
unpickled. Unfortunately unpickling data from an untrusted source is a security
risk. Therefore Listener
and Client()
use the hmac
module
to provide digest authentication.
An authentication key is a byte string which can be thought of as a password: once a connection is established both ends will demand proof that the other knows the authentication key. (Demonstrating that both ends are using the same key does not involve sending the key over the connection.)
If authentication is requested but no authentication key is specified then the
return value of current_process().authkey
is used (see
Process
). This value will be automatically inherited by
any Process
object that the current process creates.
This means that (by default) all processes of a multi-process program will share
a single authentication key which can be used when setting up connections
between themselves.
Suitable authentication keys can also be generated by using os.urandom()
.
Gerando logs¶
Some support for logging is available. Note, however, that the logging
package does not use process shared locks so it is possible (depending on the
handler type) for messages from different processes to get mixed up.
- multiprocessing.get_logger()¶
Returns the logger used by
multiprocessing
. If necessary, a new one will be created.When first created the logger has level
logging.NOTSET
and no default handler. Messages sent to this logger will not by default propagate to the root logger.Note that on Windows child processes will only inherit the level of the parent process’s logger – any other customization of the logger will not be inherited.
- multiprocessing.log_to_stderr(level=None)¶
This function performs a call to
get_logger()
but in addition to returning the logger created by get_logger, it adds a handler which sends output tosys.stderr
using format'[%(levelname)s/%(processName)s] %(message)s'
. You can modifylevelname
of the logger by passing alevel
argument.
Below is an example session with logging turned on:
>>> import multiprocessing, logging
>>> logger = multiprocessing.log_to_stderr()
>>> logger.setLevel(logging.INFO)
>>> logger.warning('doomed')
[WARNING/MainProcess] doomed
>>> m = multiprocessing.Manager()
[INFO/SyncManager-...] child process calling self.run()
[INFO/SyncManager-...] created temp directory /.../pymp-...
[INFO/SyncManager-...] manager serving at '/.../listener-...'
>>> del m
[INFO/MainProcess] sending shutdown message to manager
[INFO/SyncManager-...] manager exiting with exitcode 0
For a full table of logging levels, see the logging
module.
The multiprocessing.dummy
module¶
multiprocessing.dummy
replicates the API of multiprocessing
but is
no more than a wrapper around the threading
module.
In particular, the Pool
function provided by multiprocessing.dummy
returns an instance of ThreadPool
, which is a subclass of
Pool
that supports all the same method calls but uses a pool of
worker threads rather than worker processes.
- class multiprocessing.pool.ThreadPool([processes[, initializer[, initargs]]])¶
A thread pool object which controls a pool of worker threads to which jobs can be submitted.
ThreadPool
instances are fully interface compatible withPool
instances, and their resources must also be properly managed, either by using the pool as a context manager or by callingclose()
andterminate()
manually.processes is the number of worker threads to use. If processes is
None
then the number returned byos.process_cpu_count()
is used.If initializer is not
None
then each worker process will callinitializer(*initargs)
when it starts.Unlike
Pool
, maxtasksperchild and context cannot be provided.Nota
A
ThreadPool
shares the same interface asPool
, which is designed around a pool of processes and predates the introduction of theconcurrent.futures
module. As such, it inherits some operations that don’t make sense for a pool backed by threads, and it has its own type for representing the status of asynchronous jobs,AsyncResult
, that is not understood by any other libraries.Users should generally prefer to use
concurrent.futures.ThreadPoolExecutor
, which has a simpler interface that was designed around threads from the start, and which returnsconcurrent.futures.Future
instances that are compatible with many other libraries, includingasyncio
.
Programming guidelines¶
There are certain guidelines and idioms which should be adhered to when using
multiprocessing
.
All start methods¶
The following applies to all start methods.
Avoid shared state
As far as possible one should try to avoid shifting large amounts of data between processes.
It is probably best to stick to using queues or pipes for communication between processes rather than using the lower level synchronization primitives.
Picklability
Ensure that the arguments to the methods of proxies are picklable.
Thread safety of proxies
Do not use a proxy object from more than one thread unless you protect it with a lock.
(There is never a problem with different processes using the same proxy.)
Joining zombie processes
On POSIX when a process finishes but has not been joined it becomes a zombie. There should never be very many because each time a new process starts (or
active_children()
is called) all completed processes which have not yet been joined will be joined. Also calling a finished process’sProcess.is_alive
will join the process. Even so it is probably good practice to explicitly join all the processes that you start.
Better to inherit than pickle/unpickle
When using the spawn or forkserver start methods many types from
multiprocessing
need to be picklable so that child processes can use them. However, one should generally avoid sending shared objects to other processes using pipes or queues. Instead you should arrange the program so that a process which needs access to a shared resource created elsewhere can inherit it from an ancestor process.
Avoid terminating processes
Using the
Process.terminate
method to stop a process is liable to cause any shared resources (such as locks, semaphores, pipes and queues) currently being used by the process to become broken or unavailable to other processes.Therefore it is probably best to only consider using
Process.terminate
on processes which never use any shared resources.
Joining processes that use queues
Bear in mind that a process that has put items in a queue will wait before terminating until all the buffered items are fed by the “feeder” thread to the underlying pipe. (The child process can call the
Queue.cancel_join_thread
method of the queue to avoid this behaviour.)This means that whenever you use a queue you need to make sure that all items which have been put on the queue will eventually be removed before the process is joined. Otherwise you cannot be sure that processes which have put items on the queue will terminate. Remember also that non-daemonic processes will be joined automatically.
An example which will deadlock is the following:
from multiprocessing import Process, Queue def f(q): q.put('X' * 1000000) if __name__ == '__main__': queue = Queue() p = Process(target=f, args=(queue,)) p.start() p.join() # this deadlocks obj = queue.get()A fix here would be to swap the last two lines (or simply remove the
p.join()
line).
Explicitly pass resources to child processes
On POSIX using the fork start method, a child process can make use of a shared resource created in a parent process using a global resource. However, it is better to pass the object as an argument to the constructor for the child process.
Apart from making the code (potentially) compatible with Windows and the other start methods this also ensures that as long as the child process is still alive the object will not be garbage collected in the parent process. This might be important if some resource is freed when the object is garbage collected in the parent process.
So for instance
from multiprocessing import Process, Lock def f(): ... do something using "lock" ... if __name__ == '__main__': lock = Lock() for i in range(10): Process(target=f).start()should be rewritten as
from multiprocessing import Process, Lock def f(l): ... do something using "l" ... if __name__ == '__main__': lock = Lock() for i in range(10): Process(target=f, args=(lock,)).start()
Beware of replacing sys.stdin
with a “file like object”
multiprocessing
originally unconditionally called:os.close(sys.stdin.fileno())in the
multiprocessing.Process._bootstrap()
method — this resulted in issues with processes-in-processes. This has been changed to:sys.stdin.close() sys.stdin = open(os.open(os.devnull, os.O_RDONLY), closefd=False)Which solves the fundamental issue of processes colliding with each other resulting in a bad file descriptor error, but introduces a potential danger to applications which replace
sys.stdin()
with a “file-like object” with output buffering. This danger is that if multiple processes callclose()
on this file-like object, it could result in the same data being flushed to the object multiple times, resulting in corruption.If you write a file-like object and implement your own caching, you can make it fork-safe by storing the pid whenever you append to the cache, and discarding the cache when the pid changes. For example:
@property def cache(self): pid = os.getpid() if pid != self._pid: self._pid = pid self._cache = [] return self._cache
The spawn and forkserver start methods¶
There are a few extra restrictions which don’t apply to the fork start method.
More picklability
Ensure that all arguments to
Process.__init__()
are picklable. Also, if you subclassProcess
then make sure that instances will be picklable when theProcess.start
method is called.
Global variables
Bear in mind that if code run in a child process tries to access a global variable, then the value it sees (if any) may not be the same as the value in the parent process at the time that
Process.start
was called.However, global variables which are just module level constants cause no problems.
Safe importing of main module
Make sure that the main module can be safely imported by a new Python interpreter without causing unintended side effects (such as starting a new process).
For example, using the spawn or forkserver start method running the following module would fail with a
RuntimeError
:from multiprocessing import Process def foo(): print('hello') p = Process(target=foo) p.start()Instead one should protect the “entry point” of the program by using
if __name__ == '__main__':
as follows:from multiprocessing import Process, freeze_support, set_start_method def foo(): print('hello') if __name__ == '__main__': freeze_support() set_start_method('spawn') p = Process(target=foo) p.start()(The
freeze_support()
line can be omitted if the program will be run normally instead of frozen.)This allows the newly spawned Python interpreter to safely import the module and then run the module’s
foo()
function.Similar restrictions apply if a pool or manager is created in the main module.
Exemplos¶
Demonstration of how to create and use customized managers and proxies:
from multiprocessing import freeze_support
from multiprocessing.managers import BaseManager, BaseProxy
import operator
##
class Foo:
def f(self):
print('you called Foo.f()')
def g(self):
print('you called Foo.g()')
def _h(self):
print('you called Foo._h()')
# A simple generator function
def baz():
for i in range(10):
yield i*i
# Proxy type for generator objects
class GeneratorProxy(BaseProxy):
_exposed_ = ['__next__']
def __iter__(self):
return self
def __next__(self):
return self._callmethod('__next__')
# Function to return the operator module
def get_operator_module():
return operator
##
class MyManager(BaseManager):
pass
# register the Foo class; make `f()` and `g()` accessible via proxy
MyManager.register('Foo1', Foo)
# register the Foo class; make `g()` and `_h()` accessible via proxy
MyManager.register('Foo2', Foo, exposed=('g', '_h'))
# register the generator function baz; use `GeneratorProxy` to make proxies
MyManager.register('baz', baz, proxytype=GeneratorProxy)
# register get_operator_module(); make public functions accessible via proxy
MyManager.register('operator', get_operator_module)
##
def test():
manager = MyManager()
manager.start()
print('-' * 20)
f1 = manager.Foo1()
f1.f()
f1.g()
assert not hasattr(f1, '_h')
assert sorted(f1._exposed_) == sorted(['f', 'g'])
print('-' * 20)
f2 = manager.Foo2()
f2.g()
f2._h()
assert not hasattr(f2, 'f')
assert sorted(f2._exposed_) == sorted(['g', '_h'])
print('-' * 20)
it = manager.baz()
for i in it:
print('<%d>' % i, end=' ')
print()
print('-' * 20)
op = manager.operator()
print('op.add(23, 45) =', op.add(23, 45))
print('op.pow(2, 94) =', op.pow(2, 94))
print('op._exposed_ =', op._exposed_)
##
if __name__ == '__main__':
freeze_support()
test()
Using Pool
:
import multiprocessing
import time
import random
import sys
#
# Functions used by test code
#
def calculate(func, args):
result = func(*args)
return '%s says that %s%s = %s' % (
multiprocessing.current_process().name,
func.__name__, args, result
)
def calculatestar(args):
return calculate(*args)
def mul(a, b):
time.sleep(0.5 * random.random())
return a * b
def plus(a, b):
time.sleep(0.5 * random.random())
return a + b
def f(x):
return 1.0 / (x - 5.0)
def pow3(x):
return x ** 3
def noop(x):
pass
#
# Test code
#
def test():
PROCESSES = 4
print('Creating pool with %d processes\n' % PROCESSES)
with multiprocessing.Pool(PROCESSES) as pool:
#
# Tests
#
TASKS = [(mul, (i, 7)) for i in range(10)] + \
[(plus, (i, 8)) for i in range(10)]
results = [pool.apply_async(calculate, t) for t in TASKS]
imap_it = pool.imap(calculatestar, TASKS)
imap_unordered_it = pool.imap_unordered(calculatestar, TASKS)
print('Ordered results using pool.apply_async():')
for r in results:
print('\t', r.get())
print()
print('Ordered results using pool.imap():')
for x in imap_it:
print('\t', x)
print()
print('Unordered results using pool.imap_unordered():')
for x in imap_unordered_it:
print('\t', x)
print()
print('Ordered results using pool.map() --- will block till complete:')
for x in pool.map(calculatestar, TASKS):
print('\t', x)
print()
#
# Test error handling
#
print('Testing error handling:')
try:
print(pool.apply(f, (5,)))
except ZeroDivisionError:
print('\tGot ZeroDivisionError as expected from pool.apply()')
else:
raise AssertionError('expected ZeroDivisionError')
try:
print(pool.map(f, list(range(10))))
except ZeroDivisionError:
print('\tGot ZeroDivisionError as expected from pool.map()')
else:
raise AssertionError('expected ZeroDivisionError')
try:
print(list(pool.imap(f, list(range(10)))))
except ZeroDivisionError:
print('\tGot ZeroDivisionError as expected from list(pool.imap())')
else:
raise AssertionError('expected ZeroDivisionError')
it = pool.imap(f, list(range(10)))
for i in range(10):
try:
x = next(it)
except ZeroDivisionError:
if i == 5:
pass
except StopIteration:
break
else:
if i == 5:
raise AssertionError('expected ZeroDivisionError')
assert i == 9
print('\tGot ZeroDivisionError as expected from IMapIterator.next()')
print()
#
# Testing timeouts
#
print('Testing ApplyResult.get() with timeout:', end=' ')
res = pool.apply_async(calculate, TASKS[0])
while 1:
sys.stdout.flush()
try:
sys.stdout.write('\n\t%s' % res.get(0.02))
break
except multiprocessing.TimeoutError:
sys.stdout.write('.')
print()
print()
print('Testing IMapIterator.next() with timeout:', end=' ')
it = pool.imap(calculatestar, TASKS)
while 1:
sys.stdout.flush()
try:
sys.stdout.write('\n\t%s' % it.next(0.02))
except StopIteration:
break
except multiprocessing.TimeoutError:
sys.stdout.write('.')
print()
print()
if __name__ == '__main__':
multiprocessing.freeze_support()
test()
An example showing how to use queues to feed tasks to a collection of worker processes and collect the results:
import time
import random
from multiprocessing import Process, Queue, current_process, freeze_support
#
# Function run by worker processes
#
def worker(input, output):
for func, args in iter(input.get, 'STOP'):
result = calculate(func, args)
output.put(result)
#
# Function used to calculate result
#
def calculate(func, args):
result = func(*args)
return '%s says that %s%s = %s' % \
(current_process().name, func.__name__, args, result)
#
# Functions referenced by tasks
#
def mul(a, b):
time.sleep(0.5*random.random())
return a * b
def plus(a, b):
time.sleep(0.5*random.random())
return a + b
#
#
#
def test():
NUMBER_OF_PROCESSES = 4
TASKS1 = [(mul, (i, 7)) for i in range(20)]
TASKS2 = [(plus, (i, 8)) for i in range(10)]
# Create queues
task_queue = Queue()
done_queue = Queue()
# Submit tasks
for task in TASKS1:
task_queue.put(task)
# Start worker processes
for i in range(NUMBER_OF_PROCESSES):
Process(target=worker, args=(task_queue, done_queue)).start()
# Get and print results
print('Unordered results:')
for i in range(len(TASKS1)):
print('\t', done_queue.get())
# Add more tasks using `put()`
for task in TASKS2:
task_queue.put(task)
# Get and print some more results
for i in range(len(TASKS2)):
print('\t', done_queue.get())
# Tell child processes to stop
for i in range(NUMBER_OF_PROCESSES):
task_queue.put('STOP')
if __name__ == '__main__':
freeze_support()
test()