1. 静态化python模块

Cython提供了一些装饰器和类型用于直接扩展python模块,先看一个例子:

%%writefile logistic.py #cython: language_level=3 import cython from math import exp if cython.compiled:     print("Yep, I'm compiled.") else:     print("Just a lowly interpreted script.")  @cython.boundscheck(False) @cython.ccall def logistic(x:cython.double)->cython.double:     return 1/(1+exp(-x)) 

Writing logistic.py 

import logistic 

Just a lowly interpreted script. 

logistic.logistic(25) 

0.999999999986112 

%timeit logistic.logistic(25) 

534 ns ± 70 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each) 

!cythonize -i -3 logistic.py 

Compiling /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.py because it changed. [1/1] Cythonizing /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.py running build_ext building 'logistic' extension creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化 creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能 creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码 creating /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块 gcc -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -O2 -pipe -mcpu=arm1176jzf-s -mfpu=vfp -mfloat-abi=hard -w -I/Users/huangsizhe/LIB/CONDA/anaconda/include/python3.6m -c /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.c -o /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.o gcc -bundle -undefined dynamic_lookup -Wl,-rpath,/Users/huangsizhe/LIB/CONDA/anaconda/lib -L/Users/huangsizhe/LIB/CONDA/anaconda/lib -headerpad_max_install_names -Wl,-rpath,/Users/huangsizhe/LIB/CONDA/anaconda/lib -L/Users/huangsizhe/LIB/CONDA/anaconda/lib -headerpad_max_install_names -O2 -pipe -mcpu=arm1176jzf-s -mfpu=vfp -mfloat-abi=hard -w -arch x86_64 /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/tmp2i2j3uyu/Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.o -L/Users/huangsizhe/LIB/CONDA/anaconda/lib -o /Users/huangsizhe/WORKSPACE/Blog/Docs/Python/TutorialForPython/ipynbs/python性能优化/使用Cython优化python程序的性能/使用cython加速python源码/静态化python模块/logistic.cpython-36m-darwin.so 

import logistic 

Writing logistic.py 1

logistic.logistic(25) 

0.999999999986112 

%timeit logistic.logistic(25) 

Writing logistic.py 5

Writing logistic.py 6

Writing logistic.py 7

Writing logistic.py 8

Writing logistic.py 9

import logistic 0

Writing logistic.py 1

import logistic 2

0.999999999986112 

import logistic 4

import logistic 5

import logistic 6

import logistic 7

import logistic 8

import logistic 9

Just a lowly interpreted script. 0

Just a lowly interpreted script. 1

Just a lowly interpreted script. 2

Just a lowly interpreted script. 3

Just a lowly interpreted script. 4

Just a lowly interpreted script. 5

Just a lowly interpreted script. 6

Just a lowly interpreted script. 7

Writing logistic.py 1

Just a lowly interpreted script. 9

logistic.logistic(25) 0