Cython VS Pyston

The approach that I favour is to use Cython. The nice thing with this approach is that your code is still written as (almost) Python, but so long as you define all required types correctly it will automatically create the C extension for you. Early versions of Cython required using Cython specific typing (Python didn't have type hints when Cython was created), but it can now use Python's type hints.

and again, everything that was released after using an older version of cython.

Just for reference,

* Nuitka[0] "is a Python compiler written in Python. It's fully compatible with Python 2.6, 2.7, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 3.10, and 3.11."

* Pypy[1] "is a replacement for CPython" with builtin optimizations such as on the fly JIT compiles.

* Cython[2] "is an optimising static compiler for both the Python programming language and the extended Cython programming language... makes writing C extensions for Python as easy as Python itself."

* Numba[3] "is an open source JIT compiler that translates a subset of Python and NumPy code into fast machine code."

* Pyston[4] "is a performance-optimizing JIT for Python, and is drop-in compatible with ... CPython 3.8.12"

[0] https://github.com/Nuitka/Nuitka

[1] https://www.pypy.org/

[2] https://cython.org/

[3] https://numba.pydata.org/

[4] https://github.com/pyston/pyston

Profile and optimize the hotspots with cython (or whatever the cool kids are using these days... It's been a while.)

JIT essentially means generating machine code for the language on the fly, either during loading of the interpreter (method JIT), or by profiling and optimizing hotspots (tracing JIT). The language itself can be statically or dynamically typed. You could also compile a dynamic language ahead of time, for example, cython.

Just for reference,

* Nuitka[0] "is a Python compiler written in Python. It's fully compatible with Python 2.6, 2.7, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 3.10, and 3.11."

* Pypy[1] "is a replacement for CPython" with builtin optimizations such as on the fly JIT compiles.

* Cython[2] "is an optimising static compiler for both the Python programming language and the extended Cython programming language... makes writing C extensions for Python as easy as Python itself."

* Numba[3] "is an open source JIT compiler that translates a subset of Python and NumPy code into fast machine code."

* Pyston[4] "is a performance-optimizing JIT for Python, and is drop-in compatible with ... CPython 3.8.12"

[0] https://github.com/Nuitka/Nuitka

[1] https://www.pypy.org/

[2] https://cython.org/

[3] https://numba.pydata.org/

[4] https://github.com/pyston/pyston

and if you need more speed you can try https://www.pypy.org/ or https://github.com/tonybaloney/Pyjion or https://www.pyston.org/

What efforts have been done come with limitations. PyPy is mostly compatible. Pyston seems mostly compatible but offers only modest speedups. IronPython and Jython run on the .NET and Java runtimes, respectively. They’re JITed as a consequence of that, but that also means they’re stuck in those environments and don’t work with CPython modules that use native code.

https://github.com/pyston/pyston/blob/69b190003f14dfd2f6d276...

Seems easier to use the C functions to do this, rather than rely on system commands.

Releases · pyston/pyston

Honestly, even that seems trivial? By my reading of https://github.com/pyston/pyston#installing-packages , the only impact is that when you install (compiled) libraries they need to be recompiled, just like if you use Alpine (which is also ABI-incompatible because it uses musl libc), which is a little bit of pain at build/packaging time but doesn't actually break anything (i.e. there are no libraries that you can't use, just libraries with an extra compile step) and doesn't affect runtime behavior at all.