Generic function that applies `func` (which doesn't have to be imported in the module!) for every set of arguments in `args` and `kwargs` concurrently.

gistfile1.py

def run_multicore(func, m_args, m_kwargs, num_procs=None):
    """ Run `func` once for each set of arguments in `m_args` and `m_kwargs`, using
        either `num_procs` worker processes or as many as CPU cores are
        available.
    """
    class Worker(Process):
        def __init__(self, func, queue):
            super(Worker, self).__init__()
            self.func = func
            self.queue = queue

        def run(self):
            for params in iter(self.queue.get, None):
                self.func(*params[0], **params[1])

    if not num_procs:
        num_procs = cpu_count()

    # If args or kwargs is neither None or a list, we assume that this argument
    # is to be applied every time.
    # FIXME: There must be something in itertools that does this in a manner
    #        less verbose...
    if m_args and not isinstance(m_args, list):
        m_args = [m_args for x in xrange(len(m_kwargs or []))]
    if m_kwargs and not isinstance(m_kwargs, list):
        m_kwargs = [m_kwargs for x in xrange(len(m_args or []))]

    running = []
    queue = Queue()
    for i in xrange(num_procs):
        w = Worker(func, queue)
        running.append(w)
        w.start()
    for params in zip(m_args, m_kwargs):
        queue.put(params)
    for i in xrange(num_procs):
        queue.put(None)
    for worker in running:
        worker.join()