2024-11-16 –, LT7
Language: English
Python is infamous for its slowless and the GIL problem. But modern Python has some mechanism for parallelization, such as asyncio and concurrent.future. And free-threaded Python is coming soon. This talk will be about some design pattern on using them, focusing on using Python to generate data.
Not about the traditional multithreading or multiprocessing, which will require a lot of code to synchronize correctly. Instead, the new modules in Python made parallelization easier and less code. However, mistake can be made if you are not informed about the mechanism. I will share a few, and show the timing metrics, to show what is right. Depends on the schedule on the development of free-threaded Python into CPython, I may show how to use that as well.
(pending)