最近有個很繁瑣的需要提前計算資料指標的問題，很慢，於是就想到了並行化，之前沒有用過python進行並行化，搜了一下，受Caspar的譯文譯文啟發，原文在這裡，實現如下：

from multiprocessing import Pool

def compute(params):
    '''
    params:[param1,param2,param3...]
    '''
    # ...
    pass

def get_params(dataset):
    '''
    return the list of params, each param is also a list
    '''
 

    # ...

    params = []

    for batch_num in xrange(num_batch):
        param = []
        param.append(param1)
        param.append(param2)
        param.append(param3)

        params.append(param)

    return params

if __name__ == '__main__':


    dataset = 'xxxxxx'
    params = get_params(dataset)

    pool = Pool()
    pool.map(compute, params)
    pool.close()
    pool.join()
 

這裡用列表來儲存多個引數。