Hacker News Reader

Comment by asojfdowgh | Hacker News Reader

can't wait for this to be rather successful in hitting goals but then do nothing and languish like everything else

It seems absurd to not mention the API upstreamed by the Pyjion attempt, especially in the context of optimizing bytecode / touching into machine code?

I dunno, maybe I'm just perturbed by the amount of duplicated effort on this singular topic

View on HN

This is from the latest 3.11 alpha 4 release just this week:

The Faster CPython Project is already yielding some exciting results: this version of CPython 3.11 is ~ 19% faster on the geometric mean of the PyPerformance benchmarks, compared to 3.10.0.

https://pythoninsider.blogspot.com/2022/01/python-3102-3910-...

So this is definitely not languishing as they are getting these changes directly into main for Python 3.11

This is great, but at the same time 19% is still pretty far from 50%, and there are only ~3 months left until the feature freeze. It will be interesting to see how far they get in this release.

From following their issue tracker: some ideas didn't pan out (yet), some ideas turned out to break backwards compatibility, several pyperformance tests broke because of the changes made so they had to fix upstream (eg gevent and cython) and a lot of the work seems to be setting up future optimizations. They're also getting their changes merged directly into CPython main, so I reckon there's more red tape than if they had done the work on a fork. It does seem to inspire fellow core devs to pitch in and try some of their own ideas, so I can see the effort pick up steam as they make more headway.

Interesting, where is their issue tracker?

There you go: https://github.com/faster-cpython/ideas

Reply on news.ycombinator.com