linux-kernel - Re: [PATCH V3] KSM: allow dedup all tasks memory

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAGqmi76_ftDGtyowNRz7CCxRoJ6U3L747M=dWbRjh357w3=ZKA@mail.gmail.com>
Date:   Tue, 13 Nov 2018 14:25:45 +0300
From:   Timofey Titovets <timofey.titovets@...esis.ru>
To:     Matthew Wilcox <willy@...radead.org>
Cc:     Linux Kernel <linux-kernel@...r.kernel.org>, linux-mm@...ck.org,
        linux-doc@...r.kernel.org
Subject: Re: [PATCH V3] KSM: allow dedup all tasks memory

вт, 13 нояб. 2018 г. в 04:49, Matthew Wilcox <willy@...radead.org>:
>
> On Tue, Nov 13, 2018 at 02:13:44AM +0300, Timofey Titovets wrote:
> > Some numbers from different not madvised workloads.
> > Formulas:
> >   Percentage ratio = (pages_sharing - pages_shared)/pages_unshared
> >   Memory saved = (pages_sharing - pages_shared)*4/1024 MiB
> >   Memory used = free -h
> >
> >   * Name: My working laptop
> >     Description: Many different chrome/electron apps + KDE
> >     Ratio: 5%
> >     Saved: ~100  MiB
> >     Used:  ~2000 MiB
>
> Your _laptop_ saves 100MB of RAM?  That's extraordinary.  Essentially
> that's like getting an extra 100MB of page cache for free.  Is there
> any observable slowdown?  I could even see there being a speedup (due
> to your working set being allowed to be 5% larger)
>
> I am now a big fan of this patch and shall try to give it the review
> that it deserves.

I'm not sure if this is sarcasm,
anyway i try do my best to get that working.

On any x86 desktop with mixed load (browser, docs, games & etc)
You will always see something like 40-200 MiB of deduped pages,
based on type of load of course.

I'm just don't try use that numbers as reason to get general KSM
deduplication in kernel.
Because in current generation with several gigabytes of memory,
several saved MiB not looks serious for most of people.

Thanks!