mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Rik van Riel <riel@surriel.com>
To: Chris Li <chrisl@kernel.org>
Cc: "Gregory Price" <gourry@gourry.net>,
	"Johannes Weiner" <hannes@cmpxchg.org>,
	"Baoquan He" <baoquan.he@linux.dev>,
	"Nhat Pham" <nphamcs@gmail.com>,
	"Kairui Song" <kasong@tencent.com>,
	"Michal Hocko" <mhocko@kernel.org>,
	"Roman Gushchin" <roman.gushchin@linux.dev>,
	"Shakeel Butt" <shakeel.butt@linux.dev>,
	"Yosry Ahmed" <yosry@kernel.org>,
	"David Hildenbrand" <david@kernel.org>,
	"Muchun Song" <muchun.song@linux.dev>,
	"Kemeng Shi" <shikemeng@huaweicloud.com>,
	"Barry Song" <baohua@kernel.org>,
	"YoungJun Park" <youngjun.park@lge.com>,
	"Chengming Zhou" <chengming.zhou@linux.dev>,
	"Lorenzo Stoakes (Oracle)" <ljs@kernel.org>,
	"Liam R. Howlett" <liam@infradead.org>,
	"Vlastimil Babka (SUSE)" <vbabka@kernel.org>,
	"Mike Rapoport" <rppt@kernel.org>,
	"Suren Baghdasaryan" <surenb@google.com>,
	"Qi Zheng" <qi.zheng@linux.dev>,
	"Axel Rasmussen" <axelrasmussen@google.com>,
	"Yuanchu Xie" <yuanchu@google.com>, "Wei Xu" <weixugc@google.com>,
	"Wenchao Hao" <haowenchao22@gmail.com>,
	"Jonathan Corbet" <corbet@lwn.net>,
	"Hugh Dickins" <hughd@google.com>,
	"Baolin Wang" <baolin.wang@linux.alibaba.com>,
	"Tejun Heo" <tj@kernel.org>, "Michal Koutný" <mkoutny@suse.com>,
	"Shuah Khan" <skhan@linuxfoundation.org>,
	"Kunwu Chan" <kunwu.chan@linux.dev>,
	"Meta kernel team" <kernel-team@meta.com>,
	"Linux Memory Management List" <linux-mm@kvack.org>,
	"Linux Kernel Mailing List" <linux-kernel@vger.kernel.org>,
	linux-doc@vger.kernel.org,
	"open list:CONTROL GROUP - MEMORY RESOURCE CONTROLLER (MEMCG)"
	<cgroups@vger.kernel.org>,
	"Andrew Morton" <akpm@linux-foundation.org>,
	"Kairui Song" <ryncsn@gmail.com>,
	"Joshua Hahn" <joshua.hahnjy@gmail.com>
Subject: Re: Path forward for Virtualized Swap?
Date: Tue, 22 Sep 2026 10:59:05 -0400	[thread overview]
Message-ID: <a75b2eac4297ca6b238dde6f9246d2c1807b935f.camel@surriel.com> (raw)
In-Reply-To: <CACePvbXZQERwLF_f2FYTHsv--kq_-Aw2vVVFF+z+ovdnsHy1bA@mail.gmail.com>

On Tue, 2026-09-22 at 03:32 -1000, Chris Li wrote:
> On Mon, Sep 21, 2026 at 8:53 AM Rik van Riel <riel@surriel.com>
> wrote:
> > 
> > On Mon, 2026-09-21 at 08:10 -1000, Chris Li wrote:
> > > On Mon, Sep 21, 2026 at 6:40 AM Rik van Riel <riel@surriel.com>
> > > wrote:
> > > > 
> > > > On Mon, 2026-09-21 at 06:31 -1000, Chris Li wrote:
> > > > > 
> > > > > Give me some real-world examples of people hitting this kind
> > > > > of
> > > > > limit.
> > > > > e.g. MAX(%)  > 100% and still have a usesable system. I
> > > > > suspect
> > > > > that
> > > > > with 100% zswap usage, the system becomes unusable. Please
> > > > > show
> > > > > me a
> > > > > real world case that proves me wrong.
> > > > > 
> > > > 
> > > > Here's my desktop system right now. The amount
> > > > of Zswapped is more than twice AnonPages, and
> > > > the system stays perfectly usable.
> > > > 
> > > > $ grep -E '(Anon|Zswap|GPU|Swap)' /proc/meminfo
> > > > SwapCached:       252340 kB
> > > > SwapTotal:      41523604 kB
> > > > SwapFree:        9802088 kB
> > > > Zswap:           8812236 kB
> > > > Zswapped:       24927876 kB
> > > > AnonPages:      11615760 kB
> > > > AnonHugePages:   1402880 kB
> > > > GPUActive:      13112428 kB
> > > > GPUReclaim:        95980 kB
> > > > 
> > > > This is also an example where I cannot really
> > > > predict how much zswap space I need, because
> > > > that depends on what AI model is loaded,
> > > > which drives the GPUActive number, as well
> > > > as the number of browser tabs.
> > > > 
> > > > The system is perfectly responsive.
> > > 
> > > Great, that is getting somewhere. We are talking about actual
> > > usage.
> > > I
> > > really appreciate that.
> > > 
> > > May I ask what the total amount of memory (RAM) in your system
> > > is?
> > > 
> > > I am not asking zswap compare to ANA memory. I want to get the %
> > > of
> > > the total RAM.
> > > I assume your desktop has more than 8G of RAM, right?
> > 
> > Right now it has 64GB of memory.
> 
> That is exactly my point. You are running 1/8 = 12.5% system ram.
> 

> My point is that, I haven't seen a case where the application is
> still
> usable with a reasonable SLO when 100% of RAM is zswapped out.

In my numbers above, the applications have 11.6GB
AnonPages, and 25GB Zswap.

That is about 68% of anon memory stored in zswap. Not 8%.

Out of (anon + zswap), zswap takes up 43%

Can we at least agree on the observed numbers?
> > 

> > hen the amount needed to fit a workload
> > does not have a positive correlation with
> > the amount of system memory?
> 
> It absolutely has a positive correlation with % of the system memory
> based on my experience. The % of memory swap out has a very positive
> correlation with the thrashing on the application.

The applications thrash much harder once zswap flows
over onto disk! Out of all the swap options available,
zswap is the fastest, lowest latency one.

Forcing things out to disk "because too much zswap
is slow" could just slow things down further.

Now, there is probably some upper bound where
the zswap pool itself starts to hurt performance
because it's crowding the applications out of
memory, but that could be quite large.

-- 
All Rights Reversed.

  reply	other threads:[~2026-09-22 15:01 UTC|newest]

Thread overview: 100+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-04 21:14 Nhat Pham
2026-09-07  5:51 ` Kairui Song
2026-09-08 16:36   ` Nhat Pham
2026-09-11 16:09     ` Kairui Song
2026-09-11 16:57       ` Nhat Pham
2026-09-11 18:14         ` Kairui Song
2026-09-11 19:03           ` Nhat Pham
2026-09-12  8:47             ` Kairui Song
2026-09-14 16:49               ` Nhat Pham
2026-09-08 18:30   ` Johannes Weiner
2026-09-09 16:41     ` Nhat Pham
2026-09-09 17:47       ` Nhat Pham
2026-09-12  9:00       ` Kairui Song
2026-09-12 11:51         ` Johannes Weiner
2026-09-19  7:23           ` Chris Li
2026-09-20  1:06             ` Rik van Riel
2026-09-21  0:18               ` Chris Li
2026-09-21  0:35                 ` Rik van Riel
2026-09-21  0:53                   ` Chris Li
2026-09-14 16:10         ` Nhat Pham
2026-09-10 23:27   ` Nhat Pham
2026-09-07 11:30 ` David Hildenbrand (Arm)
2026-09-08 16:45   ` Nhat Pham
2026-09-10 10:56     ` David Hildenbrand (Arm)
2026-09-10 16:22       ` Nhat Pham
2026-09-10 17:57         ` David Hildenbrand (Arm)
2026-09-11 16:20         ` Kairui Song
2026-09-11 16:56           ` David Hildenbrand (Arm)
2026-09-14 15:00           ` Baoquan He
2026-09-19  6:50             ` Chris Li
2026-09-21 18:55               ` Nhat Pham
2026-09-22 13:38                 ` Chris Li
2026-09-22 14:43                   ` Shakeel Butt
2026-09-22 14:56                     ` Chris Li
2026-09-22 15:07                       ` Johannes Weiner
2026-09-22 15:30                         ` Chris Li
2026-09-22 15:45                           ` Shakeel Butt
2026-09-23  6:44                             ` Chris Li
2026-09-22 15:46                           ` Johannes Weiner
2026-09-22 15:28                       ` Baoquan He
2026-09-10  7:09 ` Baoquan He
2026-09-10 16:39   ` Shakeel Butt
2026-09-11 13:06     ` Baoquan He
2026-09-11 16:45       ` Shakeel Butt
2026-09-15  5:48         ` Baoquan He
2026-09-15  6:44           ` Baoquan He
2026-09-19  6:55             ` Chris Li
2026-09-19 16:15               ` Rik van Riel
2026-09-19 21:21                 ` Chris Li
2026-09-19 22:50                   ` Rik van Riel
2026-09-21  0:13                     ` Chris Li
2026-09-22 17:54                       ` Nhat Pham
2026-09-23  7:44                         ` Chris Li
2026-09-19  7:07           ` Chris Li
2026-09-10 17:03   ` Johannes Weiner
2026-09-11 12:27     ` Baoquan He
2026-09-11 16:21       ` Johannes Weiner
2026-09-19  8:45     ` Chris Li
2026-09-19 16:08       ` Gregory Price
2026-09-19 19:02         ` Chris Li
2026-09-20  1:14           ` Rik van Riel
2026-09-20  1:53           ` Gregory Price
2026-09-21  9:43             ` Chris Li
2026-09-21 12:53               ` Gregory Price
2026-09-21 16:11                 ` Chris Li
2026-09-22 17:14                   ` Rik van Riel
2026-09-22 17:21                     ` Nhat Pham
2026-09-23  7:20                       ` Chris Li
2026-09-22 17:31                     ` Johannes Weiner
2026-09-23  7:32                       ` Chris Li
2026-09-21 14:06               ` Rik van Riel
2026-09-21 16:15                 ` Chris Li
2026-09-21 15:16               ` Rik van Riel
2026-09-21 16:31                 ` Chris Li
2026-09-21 16:36                   ` Rik van Riel
2026-09-21 18:10                     ` Chris Li
2026-09-21 18:52                       ` Rik van Riel
2026-09-22 13:32                         ` Chris Li
2026-09-22 14:59                           ` Rik van Riel [this message]
2026-09-22 15:23                             ` Chris Li
2026-09-22 15:43                               ` Johannes Weiner
2026-09-23  6:10                                 ` Chris Li
2026-09-22 15:48                               ` Rik van Riel
2026-09-23  7:00                                 ` Chris Li
2026-09-22 17:32                               ` Nhat Pham
2026-09-23  7:37                                 ` Chris Li
2026-09-22 15:20                           ` Gregory Price
2026-09-22 15:43                             ` Chris Li
2026-09-22 15:52                               ` Rik van Riel
2026-09-23  7:10                                 ` Chris Li
2026-09-21 10:01           ` Kairui Song
2026-09-21 13:27             ` Gregory Price
2026-09-21 15:27               ` Rik van Riel
2026-09-21 15:48                 ` Gregory Price
2026-09-21 16:32               ` Chris Li
2026-09-21 17:02                 ` Gregory Price
2026-09-23  9:39                   ` Baoquan He
2026-09-21 18:11             ` Nhat Pham
2026-09-22 17:16           ` Johannes Weiner
2026-09-10 17:16   ` Nhat Pham

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a75b2eac4297ca6b238dde6f9246d2c1807b935f.camel@surriel.com \
    --to=riel@surriel.com \
    --cc=akpm@linux-foundation.org \
    --cc=axelrasmussen@google.com \
    --cc=baohua@kernel.org \
    --cc=baolin.wang@linux.alibaba.com \
    --cc=baoquan.he@linux.dev \
    --cc=cgroups@vger.kernel.org \
    --cc=chengming.zhou@linux.dev \
    --cc=chrisl@kernel.org \
    --cc=corbet@lwn.net \
    --cc=david@kernel.org \
    --cc=gourry@gourry.net \
    --cc=hannes@cmpxchg.org \
    --cc=haowenchao22@gmail.com \
    --cc=hughd@google.com \
    --cc=joshua.hahnjy@gmail.com \
    --cc=kasong@tencent.com \
    --cc=kernel-team@meta.com \
    --cc=kunwu.chan@linux.dev \
    --cc=liam@infradead.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=ljs@kernel.org \
    --cc=mhocko@kernel.org \
    --cc=mkoutny@suse.com \
    --cc=muchun.song@linux.dev \
    --cc=nphamcs@gmail.com \
    --cc=qi.zheng@linux.dev \
    --cc=roman.gushchin@linux.dev \
    --cc=rppt@kernel.org \
    --cc=ryncsn@gmail.com \
    --cc=shakeel.butt@linux.dev \
    --cc=shikemeng@huaweicloud.com \
    --cc=skhan@linuxfoundation.org \
    --cc=surenb@google.com \
    --cc=tj@kernel.org \
    --cc=vbabka@kernel.org \
    --cc=weixugc@google.com \
    --cc=yosry@kernel.org \
    --cc=youngjun.park@lge.com \
    --cc=yuanchu@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®