From: Yu Zhao <yuzhao@google.com>
To: Alexey Avramov <hakavlad0@seznam.cz>
Cc: 21cnbao@gmail.com, Michael@michaellarabel.com,
ak@linux.intel.com, akpm@linux-foundation.org,
aneesh.kumar@linux.ibm.com, axboe@kernel.dk,
catalin.marinas@arm.com, corbet@lwn.net,
dave.hansen@linux.intel.com, hannes@cmpxchg.org,
hdanton@sina.com, jsbarnes@google.com,
linux-arm-kernel@lists.infradead.org, linux-doc@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
mgorman@suse.de, mhocko@kernel.org, page-reclaim@google.com,
riel@surriel.com, rppt@kernel.org, torvalds@linux-foundation.org,
vbabka@suse.cz, will@kernel.org, willy@infradead.org,
x86@kernel.org, ying.huang@intel.com
Subject: Re: [PATCH v7 00/12] Multigenerational LRU Framework
Date: Sat, 12 Feb 2022 14:01:23 -0700 [thread overview]
Message-ID: <YgggI+vvtNvh3jBY@google.com> (raw)
In-Reply-To: <20220212051219.183d1baf@PC>
On Sat, Feb 12, 2022 at 05:12:19AM +0900, Alexey Avramov wrote:
> Aggressive swapping even with vm.swappiness=1 with MGLRU
> ========================================================
>
> Reading a large mmapped file leads to a super agressive swapping.
> Reducing vm.swappiness even to 1 does not have effect.
Mind explaining why you think it's "super agressive"? I assume you
expected a different behavior that would perform better. If so,
please spell it out.
> Demo: https://www.youtube.com/watch?v=J81kwJeuW58
>
> Linux 5.17-rc3, Multigenerational LRU v7,
> vm.swappiness=1, MemTotal: 11.5 GiB.
>
> $ cache-bench -r 35000 -m1 -b1 -p1 -f test20000
> Reading mmapped file (file size: 20000 MiB)
> cache-bench v0.2.0: https://github.com/hakavlad/cache-bench
Writing your own benchmark is a good exercise but fio is the standard
benchmark in this case. Please use it with --ioengine=mmap.
> Swapping started with MemAvailable=71%.
> At the end 33 GiB was swapped out when MemAvailable=60%.
>
> Is it OK?
MemAvailable is an estimate (free + page cache), and it doesn't imply
any reclaim preferences. In the worst case scenario, e.g., out of swap
space, MemAvailable *may* be reclaimed.
Here is my benchmark result with file mmap + *high* swap usage. Ram
disk was used to reduce the variance in the result (and SSD wear out
if you care). More details on additional configurations here:
https://lore.kernel.org/linux-mm/20220208081902.3550911-6-yuzhao@google.com/
Mixed workloads:
fio (buffered I/O): +13%
IOPS BW
5.17-rc3: 275k 1075MiB/s
v7: 313k 1222MiB/s
memcached (anon): +12%
Ops/sec KB/sec
5.17-rc3: 511282.72 19861.04
v7: 572408.80 22235.49
cat mmap.sh
systemctl restart memcached
swapoff -a
umount /mnt
rmmod brd
modprobe brd rd_nr=2 rd_size=56623104
mkswap /dev/ram0
swapon /dev/ram0
mkfs.ext4 /dev/ram1
mount -t ext4 /dev/ram1 /mnt
memtier_benchmark -S /var/run/memcached/memcached.sock \
-P memcache_binary -n allkeys --key-minimum=1 \
--key-maximum=50000000 --key-pattern=P:P -c 1 \
-t 36 --ratio 1:0 --pipeline 8 -d 2000
sysctl vm.overcommit_memory=1
fio -name=mglru --numjobs=36 --directory=/mnt --size=1408m \
--buffered=1 --ioengine=mmap --iodepth=128 --iodepth_batch_submit=32 \
--iodepth_batch_complete=32 --rw=randread --random_distribution=random \
--norandommap --time_based --ramp_time=10m --runtime=990m \
--group_reporting &
pid=$!
sleep 200
memcached.sock -P memcache_binary -n allkeys --key-minimum=1 \
--key-maximum=50000000 --key-pattern=R:R -c 1 -t 36 --ratio 0:1 \
--pipeline 8 --randomize --distinct-client-seed
kill -INT $pid
wait
next prev parent reply other threads:[~2022-02-12 21:01 UTC|newest]
Thread overview: 74+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-08 8:18 [PATCH v7 00/12] Multigenerational LRU Framework Yu Zhao
2022-02-08 8:18 ` [PATCH v7 01/12] mm: x86, arm64: add arch_has_hw_pte_young() Yu Zhao
2022-02-08 8:24 ` Yu Zhao
2022-02-08 10:33 ` Will Deacon
2022-02-08 8:18 ` [PATCH v7 02/12] mm: x86: add CONFIG_ARCH_HAS_NONLEAF_PMD_YOUNG Yu Zhao
2022-02-08 8:27 ` Yu Zhao
2022-02-08 8:18 ` [PATCH v7 03/12] mm/vmscan.c: refactor shrink_node() Yu Zhao
2022-02-08 8:18 ` [PATCH v7 04/12] mm: multigenerational LRU: groundwork Yu Zhao
2022-02-08 8:28 ` Yu Zhao
2022-02-10 20:41 ` Johannes Weiner
2022-02-15 9:43 ` Yu Zhao
2022-02-15 21:53 ` Johannes Weiner
2022-02-21 8:14 ` Yu Zhao
2022-02-23 21:18 ` Yu Zhao
2022-02-25 16:34 ` Minchan Kim
2022-03-03 15:29 ` Johannes Weiner
2022-03-03 19:26 ` Yu Zhao
2022-03-03 21:43 ` Johannes Weiner
2022-03-11 10:16 ` Barry Song
2022-03-11 23:45 ` Yu Zhao
2022-03-12 10:37 ` Barry Song
2022-03-12 21:11 ` Yu Zhao
2022-03-13 4:57 ` Barry Song
2022-03-14 11:11 ` Barry Song
2022-03-14 16:45 ` Yu Zhao
2022-03-14 23:38 ` Barry Song
[not found] ` <CAOUHufa9eY44QadfGTzsxa2=hEvqwahXd7Canck5Gt-N6c4UKA@mail.gmail.com>
[not found] ` <CAGsJ_4zvj5rmz7DkW-kJx+jmUT9G8muLJ9De--NZma9ey0Oavw@mail.gmail.com>
2022-03-15 10:29 ` Barry Song
2022-03-16 2:46 ` Yu Zhao
2022-03-16 4:37 ` Barry Song
2022-03-16 5:44 ` Yu Zhao
2022-03-16 6:06 ` Barry Song
2022-03-16 21:37 ` Yu Zhao
2022-02-10 21:37 ` Matthew Wilcox
2022-02-13 21:16 ` Yu Zhao
2022-02-08 8:18 ` [PATCH v7 05/12] mm: multigenerational LRU: minimal implementation Yu Zhao
2022-02-08 8:33 ` Yu Zhao
2022-02-08 16:50 ` Johannes Weiner
2022-02-10 2:53 ` Yu Zhao
2022-02-13 10:04 ` Hillf Danton
2022-02-17 0:13 ` Yu Zhao
2022-02-23 8:27 ` Huang, Ying
2022-02-23 9:36 ` Yu Zhao
2022-02-24 0:59 ` Huang, Ying
2022-02-24 1:34 ` Yu Zhao
2022-02-24 3:31 ` Huang, Ying
2022-02-24 4:09 ` Yu Zhao
2022-02-24 5:27 ` Huang, Ying
2022-02-24 5:35 ` Yu Zhao
2022-02-08 8:18 ` [PATCH v7 06/12] mm: multigenerational LRU: exploit locality in rmap Yu Zhao
2022-02-08 8:40 ` Yu Zhao
2022-02-08 8:18 ` [PATCH v7 07/12] mm: multigenerational LRU: support page table walks Yu Zhao
2022-02-08 8:39 ` Yu Zhao
2022-02-08 8:18 ` [PATCH v7 08/12] mm: multigenerational LRU: optimize multiple memcgs Yu Zhao
2022-02-08 8:18 ` [PATCH v7 09/12] mm: multigenerational LRU: runtime switch Yu Zhao
2022-02-08 8:42 ` Yu Zhao
2022-02-08 8:19 ` [PATCH v7 10/12] mm: multigenerational LRU: thrashing prevention Yu Zhao
2022-02-08 8:43 ` Yu Zhao
2022-02-08 8:19 ` [PATCH v7 11/12] mm: multigenerational LRU: debugfs interface Yu Zhao
2022-02-18 18:56 ` [page-reclaim] " David Rientjes
2022-02-08 8:19 ` [PATCH v7 12/12] mm: multigenerational LRU: documentation Yu Zhao
2022-02-08 8:44 ` Yu Zhao
2022-02-14 10:28 ` Mike Rapoport
2022-02-16 3:22 ` Yu Zhao
2022-02-21 9:01 ` Mike Rapoport
2022-02-22 1:47 ` Yu Zhao
2022-02-23 10:58 ` Mike Rapoport
2022-02-23 21:20 ` Yu Zhao
2022-02-08 10:11 ` [PATCH v7 00/12] Multigenerational LRU Framework Oleksandr Natalenko
2022-02-08 11:14 ` Michal Hocko
2022-02-08 11:23 ` Oleksandr Natalenko
2022-02-11 20:12 ` Alexey Avramov
2022-02-12 21:01 ` Yu Zhao [this message]
2022-03-03 6:06 ` Vaibhav Jain
2022-03-03 6:47 ` Yu Zhao
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=YgggI+vvtNvh3jBY@google.com \
--to=yuzhao@google.com \
--cc=21cnbao@gmail.com \
--cc=Michael@michaellarabel.com \
--cc=ak@linux.intel.com \
--cc=akpm@linux-foundation.org \
--cc=aneesh.kumar@linux.ibm.com \
--cc=axboe@kernel.dk \
--cc=catalin.marinas@arm.com \
--cc=corbet@lwn.net \
--cc=dave.hansen@linux.intel.com \
--cc=hakavlad0@seznam.cz \
--cc=hannes@cmpxchg.org \
--cc=hdanton@sina.com \
--cc=jsbarnes@google.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
--cc=mhocko@kernel.org \
--cc=page-reclaim@google.com \
--cc=riel@surriel.com \
--cc=rppt@kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=vbabka@suse.cz \
--cc=will@kernel.org \
--cc=willy@infradead.org \
--cc=x86@kernel.org \
--cc=ying.huang@intel.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;
as well as URLs for NNTP newsgroup(s).