[Python-ideas] An alternate approach to async IO

Trent Nelson trent at snakebite.org
Tue Nov 27 23:36:45 CET 2012


On Tue, Nov 27, 2012 at 02:12:12PM -0800, Sturla Molden wrote:
> 
> Den 27. nov. 2012 kl. 22:48 skrev Trent Nelson <trent at snakebite.org>:
> 
> > 
> >    Hrm, neither of those points are really flaws with my idea, you're
> >    just suggesting two different ways of working with IOCP, both of
> >    which wouldn't scale as well as a GIL-independent approach.
> 
> CPython is not a GIL-independent approach. You can stack ten thread pools om top of each other, but sooner or later you must call back to Python.

    Right, but with things like interlocked lists, you can make that
    CPython|background_IO synchronization barrier much more performant
    than relying on GIL acquisition.

    I think we'll have to agree to disagree at this point; there's not
    much point arguing further until there's some code+benchmarks on
    the table.

        Trent.



More information about the Python-ideas mailing list