Pyjion – A JIT for Python based upon CoreCLR(github.com) |
Pyjion – A JIT for Python based upon CoreCLR(github.com) |
* it keeps CPython as is.
* hence retains compat with extensions.
* but still has the potential to improve performance like pypy.
* and yet can be shipped just as an additional compiled lib inside the Python dist.
If it ever works, it will be wonderful.Looks like the authors are focused on Windows for their initial prototype development.
https://github.com/Microsoft/Pyjion/blob/master/Tests/python...
IronPython is a re-implementation of Python in C# which allows it to play nice with .NET. This is similar to Jython which is a re-implementation of Python in Java that plays nice with the JVM.
This project on the other had is not a re-implementation at all, but aims to augment the existing CPython implementation with JIT APIs and is using CoreCLR for the JIT.
A while back there was some talk of native excel-python integration. Any chance that is in the pipes, or is MS sticking to R in that sense?
It's exploratory in the sense that we are seeing if this general approach works for CPython. But if this exploratory work shows promise, then we will see it through. And since I'm also a member of the Python development team I always want to see Python be faster and succeed more. :)
Pyston seems to be catching up to PyPy fairly quickly ...not sure if there is anything transferable to pyjion, but if you haven't done so might be worth checking out.
Also for getting around the GIL, PyParallel seems to have a really interesting (and seemingly successful approach: http://pyparallel.org/
Its being worked on by one of the Guy's at continuum...and It's definitely in continuum's best interest to help python's future so I'm sure they wouldn't mind collab.
We are also aware of Pyston. There might be a chance that they could end up using our JIT hooks, but they also started a full year before us so they are not currently structured to plug into CPython like we're trying to through the API we're designing (although hopefully they will be able to, but who knows).
Once we're back into Python, though, Pyjion's JIT'd version of whatever would kick in and do its thing. Perfect combo.