There is more than one way to run a Python - program that is. The latest release of PyPy provides another alternative that is fast and interesting.

The main implementation of Python, the one most of use, is CPython and this is generally considered the language standard. PyPy is another implemenation but one written in Python itself.

To make it possible to optimise, a restricted version of Python called RPython is used and this is almost as important as PyPy itself. It is used to implement and experiment with other dynamic languages, for example HippyVM for PHP is a fast PHP implemenation in RPython. In this release the RPython documentation has been separated out from that of PyPy. 

PyPy is faster than CPython, factors of six and more are often quoted, and really its only drawback is that it only supports Python 2 code. There is a project, Py3k, to extend support to Python 2 and there is a 3.2.5 compatible version, but with some extra funding the target is to support 3.3. 

The latest version of PyPy will run on x86 and ARM hardware and Windows, Linux, Max and OpenBSB. It is claimed to run 10% faster than the previous version due to improvements in the JIT. Mixing C code in with Python is faster because C data structures can be passed as raw pointers to pinned memory allocations. 

There have also been improvments to the builtin numpy library, which are to be explained in detail in a later blog post. As numpy is the standard and important library for scientific number crunching, using it under PyPy for the extra speed is very attractive and in the new release almost everything works, as it does under CPython. 

The original PyPy project was initially funded by the EU, then Google and even the Raspberry Pi foundation chipped in with help to get an ARM version up and running. However, the project is currently unfunded.

As the announcement of the new version says:

"We’ve shown quite a bit of progress, but we’re slowly running out of funds. Please consider donating more, or even better convince your employer to donate, so we can finish those projects! "

So if you use Python for anything, consider helping out. 

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info