性能测试：Pypy vs CPython

使用 nginx 反向代理，用 fastcgi 跑 django

Python libraries known to work under PyPy (the list is not exhaustive):
ctypes
django (without any DB but sqlite)
twisted (without ssl support)
pylons
divmod's nevow
pyglet

测试环境

OS: Mac
主机: 2.4GHz Intel Core 2 Duo
内存：4GB 1067 MHz DDR3

虚拟机:

内存：512 MB
nginx version: nginx/0.7.65
flup-1.0.1
Python 2.7.1 (b590cf6de419, Apr 30 2011, 02:00:38)
[PyPy 1.5.0-alpha0 with GCC 4.4.3]

framework: Django-1.3 

django views:

因为很多库不兼容，所以只测试简单的算法。

import hashlib
import random
from django.http import HttpResponse

def index(request):
    # generate 100 stringns with 100 random characters
    string_list = {}

    for string_index in xrange(100):
        string_list[string_index] = ""
        for index in xrange(100):
            char = chr(random.randint(97, 97+25))
            string_list[string_index] += char

    html = ""
    for _string in string_list.values():
        html += "

%s

" % hashlib.md5(_string).hexdigest()

    return HttpResponse("""%s""" % html)

测试命令：

/home/waylybaye/pypy/bin/pypy manage.py runfcgi method=threaded host=127.0.0.1 port=3033
python manage.py runfcgi method=threaded host=127.0.0.1 port=3033
ab -n 1000 -c 200 

测试结果

pypy 性能明显的好于 python ，不过很多不兼容的库是一大软肋。 因为mysql-python 不兼容所以没法测试读取数据库 的性能 (可以用sqlite)

另外测试中pypy 不太稳定，测试过程中突然就死了几次。