Python Performance Enhancers  

Python, like Java, is interpreted at run time.  And like Java, Python is compiled to byte codes and saved.  With today's processors, Python's runtimes reasonably quick, but not as fast as natively compiled C/C++.  (Python and Perl runtimes are roughly equivalent.) [ Ref 1 ] 

Should perfomance still be an issue, there are several techniques you can use to improve Python's performance.  Profile your application to understand it's bottlenecks and consider one or more of the following techniques. 

• "Cython is an optimising static compiler for both the Python programming language and the extended Cython programming language.  It makes writing C extensions for Python as easy as Python itself."  Cython is based on Pyrex.

• "PyPy is an implementation of Python that makes use of a JIT and is very quick.  ...after a slight compilation delay. 

Older and lesser used techniques include:

• Psyco is a Specializing Compiler, not unlike Java's JIT.  You write Python functions as you normally would, but for those that you want to accelerate, you pass the function references to Psyco.  Psyco returns new function references you can use in lieu of the originals.  You may see speed increases of 2x, 3x, occasionally 100x.  Like Java, memory is the tradeoff. 

  Java's JIT, if on, is on full-time and applied to all code.  Psyco is not on unless you invoke the compiled functions, and you should only do that for those portions of your application that truely need the acceleration.  Pysco only works on 32-bit versions of Python. For 64-bit JIT acceleration one should look to PyPy.

• Pyrex is Python with a few additional keywords used to designate variables' data type.  The code is then saved as a .pyx file, run through Pyrex to create a .c file, and compiled with gcc or VC++.  The resulting extensions are similar to those created by handcoding C/C++ extensions. 

• SWIG simplifies the process recoding selected routines in C/C++, compiled with gcc or VC++, and used as native extensions. 

• PyInline is the Python equivalent of Perl's Inline.