Advanced Python Programming

Getting around the GIL with native extensions

There are Python native extensions that are written in C/C++, and are therefore able to avoid the limitations that the GIL sets out; one example is the most popular Python scientific computing package, NumPy. Within these extensions, manual releases of the GIL can be made, so that the execution can simply bypass the lock. However, these releases need to be implemented carefully and accompanied by the reassertion of the GIL before the execution goes back to the main Python execution.