We (as in @DrTodd13 and myself) would like to share with the community that we are working on a project called PyOMP that implements the OpenMP parallel programming model for Python using Numba.
The benefit we’re going for is to support parallelizing general HPC (or other) algorithms in Python through OpenMP’s tried and tested parallelism abstractions and execute fast using Numba’s JIT compilation.
Here’s our repo: GitHub - Python-for-HPC/PyOMP: OpenMP for Python in Numba
We support conda installation, and provide containers or a binder Jupyter to try PyOMP out. Also, we offer documentation and support on how to use PyOMP.
We would very happy to engage with interested users!
1 Like
Great to hear this project got a second life!
1 Like
PyOMP v0.2.0 — Now a Standalone Package!
I’m excited to share that PyOMP has hit a major milestone: we’ve just released v0.2.0! 
What’s new?
- Standalone package – PyOMP now depends on Numba instead of living in a fork.
- Faster updates – We can roll out improvements faster to keep pace with new Numba releases.
- Same great performance – All your favorite OpenMP-style directives in Python keep benefiting from Numba’s JIT.
Getting started
Simple one-line installation using conda
conda install -c python-for-hpc -c conda-forge pyomp
Learn more & contribute
Project repo & docs: GitHub - Python-for-HPC/PyOMP: OpenMP for Python in Numba
Thanks to contributors @DrTodd13 @stuartarchibald
2 Likes
Another update!
PyOMP v0.3.0 — Now with pip install!
I’m excited to share that PyOMP reached v0.3.0!
What’s new?
• Pip installation — you can now install PyOMP directly with pip
• Broader Numba support — compatible with versions 0.57–0.60
Install with
Pip
pip install pyomp
Conda
conda install -c python-for-hpc -c conda-forge pyomp
Thanks to contributors @DrTodd13 @stuartarchibald
1 Like