The Future of Python Threading

Cameron Laird claird at lairds.us
Fri Aug 10 16:14:26 CEST 2007


In article <5i329uF3muir4U1 at mid.individual.net>,
Bjoern Schliessmann  <usenet-mail-0306.20.chr0n0ss at spamgourmet.com> wrote:
>Justin T. wrote:
>
>> The detrimental effects of the GIL have been discussed several
>> times and nobody has ever done anything about it. 
>
>Also it has been discussed that dropping the GIL concept requires
>very fine locking mechanisms inside the interpreter to keep data
>serialised. The overhead managing those fine ones properly is not
>at all negligible, and I think it's more than managing GIL.
>
>> The truth is that the future (and present reality) of almost every
>> form of computing is multi-core,
>
>Is it? 8)
			.
			.
			.
I reinforce some of these points slightly:
A.  An effective change to the GIL impacts not just
    Python in the sense of the Reference Manual, but
    also all its extensions.  That has the potential
    to be a big, big cost; and
B.  At least part of the attention given multi-core
    processors--or, perhaps more accurately, the
    claims that multi-cores deserve threading re-
    writes--smells like vendor manipulation.



More information about the Python-list mailing list