I read the FAQ, and they answer a lot of “why not X” questions (make CPython faster, use Julia, etc), but comparison to #numba is a glaring omission. Doubtless they can’t be ignorant of numba, so it’s probably because it doesn’t look good: numba is wicked fast and the API is super easy.

Anyway, the Anaconda engineering team has a great comparison here: https://engineering.anaconda.com/2023/09/numba-mojo.html

2/3

Learn how to make your #Python code perform faster during our interactive Parallel #Programming #Workshop, and solve practical problems using #Dask, #Numba, and #Snakemake.
Register now!
https://www.esciencecenter.nl/event/parallel-programming-in-python-3/

By setting `nogil=True` on a Numba function, I just cut the runtime of one part in the code from 60 minutes to 4 minutes.

Feels like an okay win for essentially a single line edit.

#python #nogil #numba #threading

We are organizing a GPU programming workshop the 18th and 19th of April. The workshop takes place at the Netherlands eScience Center in Amsterdam. Registration is free for academics working in the Netherlands.

https://www.esciencecenter.nl/event/gpu-programming-2/

#GPU #Python #CuPy #Numba #CUDA

@jannem in my course, we use both #numba and #multiprocessing. It's amazing how easily you can get by with accelerating something way more by just throwing in a JIT compile decorator than by rewriting your code to run multiprocessing. It easily makes multiprocessing seem not worth the work.

#Barbagroup News—preprint on #arXiv: "PyExaFMM: an exercise in designing high-performance software with #Python and #Numba"
https://arxiv.org/abs/2303.08394

We find that designing performant Numba code for complex algorithms can be as challenging as writing in a compiled language.

With just ONE line, decorate a #Python function with ‘@njit’ to parallelize it automatically. Multi-core (CPu and GPU) and SIMD vectorization supported. *jaw drops* 😲

https://numba.pydata.org/

#simd #parallelism #numba #jit #decorators

The Real C++ Killers (Not You, Rust)
https://hackernoon.com/the-real-c-killers-not-you-rust

"…Rust, Julia, and D, do not address the problem of the XXI century. … They… help… write more features with fewer bugs, but… not of much help… to squeeze… very last flop from… hardware…"

"Spiral really solves an optimization problem in… mathematical sense.… defines run time as a target function and looks for its global optimum in… factor space of implementation variants limited by the hardware ."

#CPP #Spiral #Numba #Python

```
$ numba -s
[...]
CUDA Driver Version : 12.0
```

It was painful #cuda #numba #python

@elhult
Python has at least 3 JIT implementations #PyPy and #Numba, and also #Pyjion - all three take completely different approaches. I suspect this poll is more about the Pyjion approach. I think scientific computing applications would benefit more from a JIT in CPython.

Multithreading while lightweight and possibly truly parallel without a GIL would have the added drawback of working with threading locks, depending on the use case. So I am a bit wary about it.

@brettcannon @pythonbytes

Controversial opinion: I am sick of JIT compilation (#Numba #julia ) for scientific #programming

- During development, you have to recompile everything anyway, so no advantage over compiled languages.
- For interactive use, they never seem to get caching right so you end up CONSTANTLY recompiling things.

Clearly I should give up and go to F77 :P

RT @misterrios
Super excited for @cheukting_ho speaking about #numba at @PyBerlinPython! #python

Hello #fediverse! Time for an #introduction 🚀

poliastro is an #opensource (MIT) pure #Python library for interactive #Astrodynamics and #OrbitalMechanics. It provides a simple and intuitive API, it's fast thanks to #numba, and it handles physical quantities with units thanks to #Astropy.

🌐 Website https://www.poliastro.space/
:github: Source code https://github.com/poliastro/poliastro

We will be posting tips, news about the project, and cool plots. Stay tuned!

#python 3.11 et les alternatives avec JIT (#numba/#Pypy) sont toujours plus lents que #javascript.

#node.js/expressJS est plus rapide que #fastAPI

Quelle dommage alors que la syntaxe de Python est bien meilleure que celle de Javascript.

😭

#python 3.11 et les alternatives avec JIT (#numba/#Pypy) sont toujours plus lents que #javascript.

#node.js/expressJS est plus rapide que #fastAPI

Quelle dommage alors que la syntaxe de Python est bien meilleure que celle de Javascript.

😭

"My Summer at @anacondainc@twitter.com: Porting NumPy's `random` module to Numba" by Kaustubh Chaudhari #numba #numpy
https://engineering.anaconda.com/2022/08/numba.html

RT @ivanziogeo@twitter.com

Still working on #Hachures project... I've used #Python + #Numba to speed up the code. #Hachures symbology created using geometry generator and #tapering method from @underdarkGIS@twitter.com
Computation area: 1756 km²
Execution time: 48 s
#cartography #cartografia #gis #mappe #maps #qgis

🐦🔗: https://twitter.com/ivanziogeo/status/1226248487379095552