Sebastian Gray said @ 2025-12-05 07:32:10:

Hi, I ran into a similar problem after upgrading to Python 3.11 and can share a few things that helped. Python 3.11 is faster in many cases, but not all libraries are equally optimized, so slowdowns can appear, especially with Pandas and NumPy crossy road I

suggest some practical tests:

Check if the libraries you use are compiled correctly for your version of Python. Sometimes installing via pip results in a generic build that doesn't take full advantage of optimizations.

Try isolating the async functions that block the loop. Even a small amount of synchronous code inside an async function can slow down the speed.

If you're using Pandas for heavy operations, try moving the calculations to pure NumPy or use polars, which often has better performance.


Consider using multiprocessing instead of asyncio for CPU-bound operations. Asyncio works well when the limit is I/O, not computation.

This isn't a rare problem after upgrading, so with some targeted testing, you should be able to find the problem. If you share a small example of the code I can try to look at it better.