[Baypiggies] Web Crawler/Backend Engineer - San Francisco, CA

[Alec Flett]

Or even kind of CPU-bound, as I discovered some months back... lets say you
have a process that would normally be about 50% blocked on IO, and 50% of
the time chewing on the data retrieved from that IO blocking. You run more
than 2-3 of those on a multi-processor box, and you start getting
contention.

desperately hoping some of the various things being tried in unladen-swallow
pay off,

Alec

On Tue, Feb 2, 2010 at 2:39 PM, Asher Langton <langton2 at llnl.gov> wrote:

> On Feb 2, 2010, at 2:14 PM, Aahz wrote:
>
>> On Tue, Feb 02, 2010, Tung Wai Yip wrote:
>>
>>>
>>> I think my experience tells me don't build multi-threaded applications in
>>> Python ;-)
>>>
>>
>> Why?  I've had about equal trouble with multi-processing and
>> multi-threading.
>>
>
> If you haven't seen Dave Beazley's presentations on the Global Interpreter
> Lock, take a look at these:
>
> http://www.dabeaz.com/python/GIL.pdf
> http://www.dabeaz.com/python/NewGIL.pdf
>
> In short: multi-threading is really bad in Python 2.x when the threads are
> CPU-bound.
>
>
> -Asher
>
> _______________________________________________
> Baypiggies mailing list
> Baypiggies at python.org
> To change your subscription options or unsubscribe:
> http://mail.python.org/mailman/listinfo/baypiggies
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/baypiggies/attachments/20100202/5749c341/attachment-0001.htm>