python 小知识点

１关于joblib 的 Parallel, delayed

from joblib import Parallel, delayed
import time
pred=open("predict.csv",'wb')
train=open("train.csv","wb")
ll1=[]
ll2=[]


def func1(i):
    return i
t1=time.time()
with open("predict_first.csv","rb") as f:
    ll1=[Parallel(8)(delayed(func1)(i) for i in f)]

t2=time.time()

print(t2-t1)

t3=time.time()

with open("predict_first.csv","rb") as f:
    for i in f:
        ll2.append(func1(i))
t4=time.time()
print(t4-t3)

运行结果：

0.3481941223144531
0.03202414512634277

感觉这个并行咋提升不大，返回还慢了．