Message 379644 - 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.

In-reply-to
Author	methane
Recipients	Mark.Shannon, carljm, corona10, dino.viehland, eelizondo, gregory.p.smith, methane, nascheme, pablogsal, pitrou, remi.lapeyre, shihai1991, steve.dower, tim.peters, vstinner
Date	2020年10月26日.06:48:45
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1603694926.16.0.790801102814.issue40255@roundup.psfhosted.org>

Content
I'm big -1 too. But I am interested in Instagram usage. * How % of heap are CoW-ed with gc.freeze()? * When CoW happen? in execution time, or shutdown? * Which type cause CoW? I have two ideas to reduce CoW: * Fast shutdown option Currently Python try to free object as possible in finalization. It will cause much CoW. It will very bad for applications using multiprocessing.Pool for parallel processing of large data. * Use "legacy" Unicode representation for large strings. Unicode is cold, immutable. But refcount is not immutable. Currentlly, most unicode object is "compact": object header and data is in one memory block. We may be able to use "legacy" representation and put cold data into another page. For long term, multi process friendly `pyc` format can share more memory inter processes, even though they are not created by prefork.

Content

I'm big -1 too. But I am interested in Instagram usage.
* How % of heap are CoW-ed with gc.freeze()?
* When CoW happen? in execution time, or shutdown?
* Which type cause CoW?
I have two ideas to reduce CoW:
* Fast shutdown option
 Currently Python try to free object as possible in finalization. It will cause much CoW. It will very bad for applications using multiprocessing.Pool for parallel processing of large data.
* Use "legacy" Unicode representation for large strings.
 Unicode is cold, immutable. But refcount is not immutable. Currentlly, most unicode object is "compact": object header and data is in one memory block. We may be able to use "legacy" representation and put cold data into another page.
For long term, multi process friendly `pyc` format can share more memory inter processes, even though they are not created by prefork.

History
Date	User	Action	Args
2020年10月26日 06:48:47	methane	set	recipients: + methane, tim.peters, nascheme, gregory.p.smith, pitrou, vstinner, carljm, dino.viehland, Mark.Shannon, steve.dower, corona10, pablogsal, eelizondo, remi.lapeyre, shihai1991
2020年10月26日 06:48:46	methane	set	messageid: <1603694926.16.0.790801102814.issue40255@roundup.psfhosted.org>
2020年10月26日 06:48:46	methane	link	issue40255 messages
2020年10月26日 06:48:45	methane	create

homepage