what's the point of rpython?

Paul Rubin http
Wed Jan 21 06:37:44 CET 2009


"Rhodri James" <rhodri at wildebst.demon.co.uk> writes:
> I tend to live in single-core worlds, so "inc" on its lonesome works
> just fine.

In a single core world we wouldn't be having these endless angsty
conversations about eliminating the GIL.

> > That has already been tried, and found to be unacceptably slow for the
> > purpose at hand.  Now we're looking for the optimizations.
> 
> In that case I'd second the suggestion of taking a long, hard look
> at the Linux core locking and synchronisation primatives.

Do you understand what the issue is, about CPython's reference counts?
Do you have any idea how often the interpreter updates them?  Even
using LOCK INCR (raw machine instruction, not portable C, no operating
system involvement), manipulating the ref counts would be around 100x
slower than it is now.  



More information about the Python-list mailing list