Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

gh-115999: Update whats news for free-threaded 3.14 #131285

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions Doc/whatsnew/3.14.rst
Original file line number Diff line number Diff line change
Expand Up @@ -68,6 +68,7 @@ Summary -- release highlights
* :ref:`PEP 741: Python Configuration C API <whatsnew314-pep741>`
* :ref:`PEP 761: Discontinuation of PGP signatures <whatsnew314-pep761>`
* :ref:`A new type of interpreter <whatsnew314-tail-call>`
* :ref:`Free-threaded CPython <whatsnew314-free-threaded-cpython>`


Incompatible changes
Expand Down Expand Up @@ -311,6 +312,24 @@ For further information on how to build Python, see
(Contributed by Ken Jin in :gh:`128563`, with ideas on how to implement this
in CPython by Mark Shannon, Garrett Gu, Haoran Xu, and Josh Haberman.)

.. _whatsnew314-free-threaded-cpython:

Free-threaded CPython
---------------------

Free-threaded CPython now supports :pep:`659` which is known
as the specializing adaptive interpreter, the execution mechanisms is almost same
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
as the specializing adaptive interpreter, the execution mechanisms is almost same
as the specializing adaptive interpreter. The execution mechanism is almost same

as the default build, but at free-threaded CPython build, each thread maintains
a thread-local copy of specialized bytecode.
Comment on lines +322 to +323
Copy link
Contributor

@StanFromIreland StanFromIreland Mar 15, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
as the default build, but at free-threaded CPython build, each thread maintains
a thread-local copy of specialized bytecode.
as the default build, however in a free-threaded CPython build, each thread
maintains a thread-local copy of specialized bytecode.


With this change, the free-threaded CPython build achieve better performance
and 10-15% slower than the default build.
Comment on lines +325 to +326
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
With this change, the free-threaded CPython build achieve better performance
and 10-15% slower than the default build.
With this change, the free-threaded CPython build achieves better performance,
and is only 10-15% slower than the default build.

Maybe take out the slower part altogether ;-)


This work was done by many contributors, Matt Page, Neil Schemenauer,
Sam Gross, Thomas Wouters, Donghee Na, Kirill Podoprigora and Ken Jin.
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I also added the @Fidget-Spinner, because I remember his early design involvation while he worked at Quansight.

Copy link
Member

@Fidget-Spinner Fidget-Spinner Mar 15, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the main thing I did that is coming in 3.14 is the stackrefs. But we needed that for the intepreter anyways so it probably counts? Either way I can be left out if the others feel inclined to.

Meta also provided significant engineering resources for performance benchmark and
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Meta also provided significant engineering resources for performance benchmark and
Meta also provided significant engineering resources for performance benchmarking and

I presume this is intended?

design per thread specialized bytecode mechanisms.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
design per thread specialized bytecode mechanisms.
design of per thread specialized bytecode mechanisms.



Other language changes
======================
Expand Down
Loading