And for further information: http://jessenoller.com/2009/02/01/python-threads-and-the-global-interpreter-lock/ http://jessenoller.com/2009/02/01/python-threads-

And for further information: http://jessenoller.com/2009/02/01/python-threads-and-the-global-interpreter-lock/ http://jessenoller.com/2009/02/01/python-threads-and-the-glo... Ultimately, while the implementation of the GIL is inefficient, and does harm you in the case of processor-bound computation, they're still useful in many cases where you're dealing with I/O (file, and socket). Even as the multiprocessing maintainer, I'd like to see a more robust threading model within python - unladen swallow (http://code.google.com/p/unladen-swallow/ http://code.google.com/p/unladen-swallow/) may be able to offer us a world without a GIL soon(ish), but even barring that, getting some work done on improving what we have today would be nice.