最近工作中有個任務，就是優化一個模型的實時性。從有到無，主要完成了以下內容。

0.模型的邏輯

1.演算法邏輯

2.程式碼重構

3.程式的效能優化，包括編譯、多執行緒、多程序、numba

4.語言

numba包，經測試，比較適用於陣列、矩陣等數值計算，其他的型別操作，容易報錯。

from multiprocessing import Pool
from functools import  partial

def math_get(a, b):
    print(a)
    return b + a
def processing(b):
    pool = Pool(3)
    a = [i for i in range(3)]
    d = pool.map(partial(math_get,b = b), a,)
#多個引數代入的時候，採用partial的方式，
#其中a作為一個可迭代的物件通過map代入math_get函式
    print(d)
    pool.close()
    pool.join()
if __name__ == "__main__":
    processing(5)
 

#!/usr/bin/env python
# -*- coding:utf-8 -*-
__author__ = 'Great'
from multiprocessing import Pool
import time
def a(num):
    list_num = []
    print(num)
    for i in num:
        list_num.append(i)
        print(i)
        print(list_num)
    return list_num

if __name__ == "__main__":

    start = time.time()
    pool = Pool(3)#程序池
    num = [1,2,3,4,5,6]
   #pool.apply_async(a,args=(num,))
    a = pool.map(a,(num, ))
    print('a',a)
    pool.close()
    pool.join()
    end = time.time()
    print(end-start)
 

#!/usr/bin/env python
# -*- coding:utf-8 -*-
__author__ = 'Great'
from multiprocessing import Pool
import time
from functools import  partial
def a(num, ccc):
    list_num = []

    for i in num:
        #list_num.append(i)
        print(i + ccc)
        #print(list_num)
    #return list_num

if __name__ == "__main__":

    start = time.time()
    pool = Pool(3)

    num = [1,2,3,4,5,6]
    ccc = 1
   #for i in range(10):
   #pool.apply_async(a,args=(num,))
    pool.map(partial(a,ccc = ccc),(num,))#多個引數
    #print('a',a)
    pool.close()
    pool.join()
    end = time.time()
    print(end-start)