Re: Page replacement algorithm in buffer cache - Mailing list pgsql-hackers

From Atri Sharma
Subject Re: Page replacement algorithm in buffer cache
Date
Msg-id CAOeZVifM2-Jq_HJj268F2CLHSQcO8A-Udt_s8VFG2S=9_LniNQ@mail.gmail.com
Whole thread Raw
In response to Page replacement algorithm in buffer cache  (Atri Sharma <atri.jiit@gmail.com>)
List pgsql-hackers
On Fri, Mar 22, 2013 at 4:53 PM, Amit Kapila <amit.kapila@huawei.com> wrote:
> On Friday, March 22, 2013 4:36 PM Atri Sharma wrote:
>> >
>> > What would you do if the only young page has usage count zero during
>> second
>> > sweep.
>>
>> Umm....The same approach we take when there is no page with usage
>> count zero in a sweep in the current algorithm?
>
> It would give more priority to young page as compare to more used page.
> I don't know if that would be correct thing to do.
This is my idea, give equal priority to new pages when they enter the
cache, so that they all have an equal chance to be replaced initially.
With time, usage_count shall become the deciding factor in victim
selection.

Regards,

Atri



--
Regards,

Atri
l'apprenant



pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Page replacement algorithm in buffer cache
Next
From: Simon Riggs
Date:
Subject: Re: Should commit_delay be PGC_SIGHUP?