09-08, 14:40–15:15 (Europe/Lisbon), Auditorium
Python's flexible and intuitive syntax enables developers to quickly build applications. But on the other hand, it may be slow during runtime. Luckily, there are different ways we can speed up a Python program. In this talk, we'll explore different alternatives to make Python programs faster.
About the talk
Python's "speed problem" is not a new issue, but there are different ways to speed a Python program up. In recent years, Python 3.11 was released and popularized as the "faster Python" and Mojo programming language was recently announced, advertised as having "usability of Python with the performance of C". Going back further, we see languages such as Cython, Just In Time (JIT) compilers and bindings.
In this talk, we'll go over the different approaches to increasing the speed of a Python application. We'll briefly explain how they work, compare the performance through a simple use case, and look at the limitations, tooling, trade-offs, and ease of use.
Outline
- Introduction and setup (3min)
- Baseline - Python 3.9 (3min)
- Alternatives (20min):
- Python 3.11
- Cython & Mypyc
- Pypy3
- PyO3
- Mojo
- Recap and takeaways (4min)
Murilo is a tech lead AI at Dataroots. He takes a pragmatic approach aiming to make AI both impactful and accessible. To that end, Murilo has an emphasis on MLOps when building ML systems. He's passionate about open-source, programming, and knowledge sharing.