Message 131527 - Python tracker

➜

This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author	pitrou
Recipients	brian.curtin, kristjan.jonsson, loewis, pitrou, sbt, tim.golden
Date	2011-03-20.19:07:30
SpamBayes Score	8.303293e-08
Marked as misclassified	No
Message-id	<1300648047.3481.22.camel@localhost.localdomain>
In-reply-to	<1300647091.7.0.855144541193.issue11618@psf.upfronthosting.co.za>

Content
> Have you tried benchmarking it? > > Interlocked functions are much faster than Win32 mutex/semaphores in > the uncontended case. Well, I'd rather have obviously correct code than difficult-to-understand "speedy" code. The patch I've posted takes less than a microsecond per acquire/release pair, and that's in a virtual machine to begin with. > Do you really want the GIL to be 10 times slower in the uncontended case? ;-) The GIL doesn't use these functions (see ceval_gil.h).

> Have you tried benchmarking it?
> 
> Interlocked functions are *much* faster than Win32 mutex/semaphores in
> the uncontended case.  

Well, I'd rather have obviously correct code than
difficult-to-understand "speedy" code.

The patch I've posted takes less than a microsecond per acquire/release
pair, and that's in a virtual machine to begin with.

> Do you really want the GIL to be 10 times slower in the uncontended case? ;-)

The GIL doesn't use these functions (see ceval_gil.h).

History
Date	User	Action	Args
2011-03-20 19:07:32	pitrou	set	recipients: + pitrou, loewis, kristjan.jonsson, tim.golden, brian.curtin, sbt
2011-03-20 19:07:30	pitrou	link	issue11618 messages
2011-03-20 19:07:30	pitrou	create