From: Roman Gushchin <klamm@yandex-team.ru>
To: peterz@infradead.org, mingo@redhat.com,
Kirill Tkhai <tkhai@yandex.ru>,
LKML <linux-kernel@vger.kernel.org>
Cc: Stanislav Fomichev <stfomichev@yandex-team.ru>
Subject: [RFC] smt-aware rt load balancer
Date: Thu, 04 Sep 2014 16:20:06 +0400 [thread overview]
Message-ID: <77161409833206@webcorp01h.yandex-team.ru> (raw)
Hello!
We had an earlier discussion about using real-time policies on modern CPUs
for cpu-bound tasks with "near real-time" execution time expectations (like front-end servers):
https://lkml.org/lkml/2014/4/24/602 .
I sad, that I had a prototype of real-time load balancer (called smart), that performs well in this case.
Now it's ready to be published.
The patch set on top of the 3.10.x branch can be found here:
https://github.com/yandex/smart .
It's stable.
We use them in production for a couple of months on more than thousand machines.
We get noticeable performance increase for many projects with different load patterns
(up to 10-15% in both RPS and latency).
Any feedback, comments, questions are welcome!
Regards,
Roman
next reply other threads:[~2014-09-04 12:25 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-09-04 12:20 Roman Gushchin [this message]
2014-09-04 15:40 ` Peter Zijlstra
2014-09-04 16:30 ` [PATCH 01/19] smart: define and build per-core data structures klamm
2014-09-04 16:30 ` [PATCH 02/19] smart: add config option for SMART-related code klamm
2014-09-04 16:30 ` [PATCH 03/19] smart: introduce smart_enabled() klamm
2014-09-04 16:30 ` [PATCH 04/19] smart: helper functions for smart_data manipulations klamm
2014-09-04 16:30 ` [PATCH 05/19] smart: CPU selection logic klamm
2014-09-04 16:30 ` [PATCH 06/19] smart: use CPU selection logic if smart is enabled klamm
2014-09-21 17:31 ` Pavel Machek
2014-09-04 16:30 ` [PATCH 07/19] smart: balance load between nodes klamm
2014-09-04 16:30 ` [PATCH 08/19] smart: smart pull klamm
2014-09-04 16:30 ` [PATCH 09/19] smart: throttle CFS tasks by affinning to first SMT thread klamm
2014-09-04 16:30 ` [PATCH 10/19] smart: smart gathering klamm
2014-09-04 16:30 ` [PATCH 11/19] smart: smart debug klamm
2014-09-04 16:30 ` [PATCH 12/19] smart: cgroup interface for smart klamm
2014-09-04 16:30 ` [PATCH 13/19] smart: nosmart boot option klamm
2014-09-04 16:31 ` [PATCH 14/19] smart: smart-related sysctl's klamm
2014-09-04 16:31 ` [PATCH 15/19] smart: decrease default rt_period/rt_runtime values klamm
2014-09-04 16:31 ` [PATCH 16/19] smart: change default RR timeslice to 5ms klamm
2014-09-04 16:31 ` [PATCH 17/19] smart: disable RT runtime sharing klamm
2014-09-04 16:31 ` [PATCH 18/19] hrtimers: calculate expires_next after all timers are executed klamm
2014-09-04 16:31 ` [PATCH 19/19] smart: add description to the README file klamm
2014-09-04 16:49 ` [RFC] smt-aware rt load balancer Peter Zijlstra
2014-09-04 17:06 ` Roman Gushchin
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=77161409833206@webcorp01h.yandex-team.ru \
--to=klamm@yandex-team.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=stfomichev@yandex-team.ru \
--cc=tkhai@yandex.ru \
/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®