All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andi Kleen <andi@firstfloor.org>
To: "Steinar H. Gunderson" <sgunderson@bigfoot.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: Slowdown due to threads bouncing between HT cores
Date: Sat, 04 Oct 2014 06:41:15 -0700	[thread overview]
Message-ID: <87a95cq7es.fsf@tassilo.jf.intel.com> (raw)
In-Reply-To: <20141003194428.GA27084@sesse.net> (Steinar H. Gunderson's message of "Fri, 3 Oct 2014 21:44:29 +0200")

"Steinar H. Gunderson" <sgunderson@bigfoot.com> writes:
>
> So, benchmark:
>
>  - Default: 13266 kN/sec
>  - Change from ondemand to performance on all cores: 14600 kN/sec
>  - taskset -c 0-19 (locking affinity to only one set of hyperthreads):
>    17512 kN/sec

iirc the main reasons for unbound threads migrating away are:

- something else gets scheduled on these logical CPUs, so
the scheduler tries to balance to run queue lengths

You could check that with perf timechart or perf sched record/map
or kernelshark.

- there is some IO or communication which causes wakeup affinity.

You could try disabling WAKEUP_PREEMPTION or NEXT_BUDDY in 
/sys/kernel/debug/sched_features

-Andi
-- 
ak@linux.intel.com -- Speaking for myself only

  parent reply	other threads:[~2014-10-04 13:41 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-10-03 19:44 Slowdown due to threads bouncing between HT cores Steinar H. Gunderson
2014-10-03 21:11 ` Marc Burkhardt
2014-10-03 21:14   ` Steinar H. Gunderson
2014-10-04  9:22     ` Marc Burkhardt
2014-10-04 13:41 ` Andi Kleen [this message]
2014-10-04 14:12   ` Steinar H. Gunderson
2014-10-04 14:50 ` Chuck Ebbert
2014-10-05 11:19   ` Steinar H. Gunderson
2014-10-08 15:37 ` bisected: futex regression >= 3.14 - was - " Mike Galbraith
2014-10-08 16:14   ` Thomas Gleixner
2014-10-08 16:45     ` Steinar H. Gunderson
2014-10-08 17:52       ` Mike Galbraith
2014-10-08 16:23   ` Steinar H. Gunderson
2014-10-08 17:04   ` Linus Torvalds
2014-10-08 17:05     ` Steinar H. Gunderson
2014-10-08 17:59     ` Mike Galbraith
2014-10-24 15:25       ` Thomas Gleixner
2014-10-24 16:38         ` Mike Galbraith
2014-10-26 10:39           ` Steinar H. Gunderson
2014-10-26 13:16             ` Mike Galbraith
2014-10-26 13:58               ` Mike Galbraith
2014-10-26 14:11                 ` Steinar H. Gunderson
2014-10-26 14:41                   ` Mike Galbraith
2014-10-27 10:05                   ` Mike Galbraith

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=87a95cq7es.fsf@tassilo.jf.intel.com \
    --to=andi@firstfloor.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sgunderson@bigfoot.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.