Pyevolve benchmark on different Python flavors

So I did a comparative of Pyevolve GP/GA core in different Python interpreters. I’ve used my Pentium Core 2 Duo (E4500 @ 2.20GHz, 1GB RAM), using Ubuntu 9.04 and Windows XP SP3 just for IronPython 2.6.1 (IronPython doesn’t run with Mono, so I used the win xp with .net 2.0).

The interpreters used were:

Unladen Swallow 2009Q2

I tried using 2009Q3 (the currently main trunk), but I think it’s unstable yet, cause it was more slow than 2009Q2, so I used 2009Q2; I compiled it with GCC 4.3.3 just using the default configure parameters (./configure).

CPython 2.6.2

I used the default CPython package of Ubuntu 9.04.

CPython 2.5.4

I used the default CPython package of Ubuntu 9.04 too, the python2.5 package.

PyPy 1.1.0 (svn:r68612)

I used the last svn version of the repository, the release 68612. My Pentium Core 2 Duo had only 1GB of RAM, and the PyPy translation process eats more RAM than Java (sorry for the joke), so I used a notebook with 3GB of RAM to create the pypy-c, what took 1 hour (I used –opt=3) and a beautiful ascii Mandelbrot fractal !

Jython 2.5.1

I used the default installer from the Jython project site. I used the Sun JRE 1.6.0_16.

IronPython 2.6.10920.0

I’ve used the 2.6 RC1 available at IronPython project site with MS .NET 2.0.

To test the GA core I’ve used this source-code (a simple sphere function):

from pyevolve import G1DList
from pyevolve import Mutators, Initializators
from pyevolve import GSimpleGA, Consts

# This is the Sphere Function
def sphere(xlist):
   total = 0
   for i in xlist:
      total += i**2
   return total

def run_main():
   genome = G1DList.G1DList(140)
   genome.setParams(rangemin=-5.12, rangemax=5.13)

   ga = GSimpleGA.GSimpleGA(genome, seed=666)

   best = ga.bestIndividual()

if __name__ == "__main__":

And to test the GP core, I’ve used this source-code (a simple symbolic regression):

from pyevolve import GTree
from pyevolve import Mutators
from pyevolve import GSimpleGA, Consts, Util
import math

rmse_accum = Util.ErrorAccumulator()

def gp_add(a, b): return a+b
def gp_sub(a, b): return a-b
def gp_mul(a, b): return a*b
def gp_sqrt(a):   return math.sqrt(abs(a))

def eval_func(chromosome):
   global rmse_accum
   code_comp = chromosome.getCompiledCode()

   for a in xrange(0, 10):
      for b in xrange(0, 10):
         evaluated     = eval(code_comp)
         target        = math.sqrt((a*a)+(b*b))
         rmse_accum   += (target, evaluated)
   return rmse_accum.getRMSE()

def main_run():
   genome = GTree.GTreeGP()
   genome.setParams(max_depth=4, method="ramped")
   genome.evaluator += eval_func

   ga = GSimpleGA.GSimpleGA(genome, seed=666)
   ga.setParams(gp_terminals       = ['a', 'b'],
                gp_function_prefix = "gp")


   best = ga.bestIndividual()

if __name__ == "__main__":

UPDATE 19/08: the x-axis is measured in “seconds“, and the y-axis is the python flavor;

The results are are described in the graph below:

pyevolve_pyvmsAs we can see, Unladen Swallow 2009Q2 did a little better performance than CPython 2.6.2, but Jython and PyPy (experimental) were left behind in that scenario, even behind IronPython 2.6.1.

Jinja2 in a Java JSP

This is a simple trick possible using Jython; to call jinja2 template engine under JSP we instantiate the PythonInterpreter class of Jython, set some parameters in the interpreter and then call jinja2 to render a template and write to the Java “out” object.

To install Jython, just download the last stable version Jython 2.5 and then install it as “Standalone” version; the installer will create a simple jython.jar file under the installation directory, copy this package to your Java web project under the \WEB-INF\lib or where you put your web application libraries.

Later, copy the jinja2 module to the \build\classes.

Create a simple template under the WebContent\templates\template.html with contents below:

This is just a Jinja2 template test !

Parameter "p" = {{ request.getParameter("p") }}

Getting session attribute: {{ session.getAttribute("session_attribute") }}

Iterating over Java array:
    {% for user in users %}
  • {{ user }}
  • {% endfor %}

And now create a file called index.jsp in the root of the WebContent, with the contents:

<%@ page language="java" contentType="text/html; charset=ISO-8859-1"
<%@page import="org.python.util.PythonInterpreter" %>

Jinja2 Template under JSP

	PythonInterpreter py = new PythonInterpreter();

	String[] users = {"User Number One", "User Number Two"};
	session.setAttribute("session_attribute", "Testing Session");
	py.set("out",     out);
	py.set("request", request);
	py.set("session", session);
	py.set("users",   users);
	py.set("context_path", application.getRealPath("/") + "templates");
	py.exec("from jinja2 import Environment, FileSystemLoader");
	py.exec("env = Environment(loader=FileSystemLoader(context_path))");
	py.exec("template = env.get_template('template.html')");

The structure will be like this:


And the output will be something like this:


Good news: IronPython and Jython compatibility !

Good news, I’ve tested Pyevolve 0.5 with the IronPython 2.0.1 and Jython 2.5b1, and the framework is working perfect (besides the performance loss when comparing with CPython 2.5 and 2.6). The only feature that do not worked is the dump of statistics with the sqlite3, sine there is no sqlite3 on the default install of Jython and IronPython. To Pyevolve work on IronPython, you must install the zlib module for IronPython and import your Python modules like this:

import sys

With this compatibility issue, you can use Pyevolve on your .NET platform or in your Java applications.