archive mirror
 help / color / mirror / Atom feed
From: Con Kolivas <>
To: Sean Neakums <>, Andrew Morton <>
	Ingo Molnar <>
Subject: Re: 2.6.0-test1-mm1
Date: Wed, 16 Jul 2003 19:15:29 +1000	[thread overview]
Message-ID: <> (raw)
In-Reply-To: <>

On Wed, 16 Jul 2003 18:07, Sean Neakums wrote:
> Andrew Morton <> writes:
> > . Another interactivity patch from Con.  Feedback is needed on this
> >   please - we cannot make much progress on this fairly subjective work
> >   without lots of people telling us how it is working for them.
> This patch seems to mostly cure an oddity I've been seeing since
> 2.5.7x, or maybe very late 2.5.6x (I forget exactly when) where
> running 'ps aux' or 'ls -l' in an xterm (and only xterm it seems; I've
> tried rxvt and aterm) would more often than not result in a wallclock
> run time of up to two seconds, instead of the usual tenth of a second
> or so, with system and user time remaining constant.  If I keep
> running 'ps aux' its output does start to become slow again, snapping
> back to full speed after a few more runs.  Kind of an odd one.

In fact I've seen a number of apps display this problem. Basically how I 
understood it was the child was spawned with a lower dynamic priority (50% of 
the bonus of the parent with the child penalty at 50%). The parent then was 
spinning madly waiting for the child but in the process it was continually 
preempting the child. Eventually the parent was seen as a cpu hog, it's 
dynamic priority dropped and the child finally got to run along side it. The 
next time you ran the same parent/child combination the problem had gone 
because the parent was at a lower dynamic priority. Changing child penalty 
back up to 95% cures this problem (but can bring other problems with it 
unless workarounds are added). While this seems more an anomaly in the coding 
of the application brought out by the scheduler, it seems to be very common.


  parent reply	other threads:[~2003-07-16  8:58 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-07-16  5:56 2.6.0-test1-mm1 Andrew Morton
2003-07-16  6:16 ` 2.6.0-test1-mm1 Joshua Kwan
2003-07-16  6:22   ` 2.6.0-test1-mm1 Andrew Morton
2003-07-16  8:07 ` 2.6.0-test1-mm1 Sean Neakums
2003-07-16  8:55   ` 2.6.0-test1-mm1 Ingo Molnar
2003-07-16 10:19     ` 2.6.0-test1-mm1 Antonio Vargas
2003-07-16 13:37       ` 2.6.0-test1-mm1 Helge Hafting
2003-07-16 14:24         ` 2.6.0-test1-mm1 Antonio Vargas
2003-07-16  9:15   ` Con Kolivas [this message]
2003-07-16 10:44 ` 2.6.0-test1-mm1 Barry K. Nathan
2003-07-16 10:58   ` 2.6.0-test1-mm1 Andrew Morton
2003-07-16 12:24     ` 2.6.0-test1-mm1 William Lee Irwin III
2003-07-16 14:32       ` 2.6.0-test1-mm1 Barry K. Nathan
2003-07-16 14:41         ` 2.6.0-test1-mm1 William Lee Irwin III
2003-07-16 15:02           ` 2.6.0-test1-mm1 Barry K. Nathan
2003-07-16 15:09             ` 2.6.0-test1-mm1 William Lee Irwin III
2003-07-16 16:38               ` 2.6.0-test1-mm1 Andrew Morton
2003-07-16 14:03     ` 2.6.0-test1-mm1 Barry K. Nathan
2003-07-16 12:53 ` 2.6.0-test1-mm1 Philippe Gramoullé
2003-07-16 17:21 ` 2.6.0-test1-mm1 Ramón Rey Vicente󮠒
2003-07-16 19:01   ` 2.6.0-test1-mm1 Zwane Mwaikambo
2003-07-16 23:25     ` 2.6.0-test1-mm1 Ramón Rey Vicente󮠒
2003-07-16 23:56       ` 2.6.0-test1-mm1 Zwane Mwaikambo
2003-07-16 22:02 ` 2.6.0-test1-mm1 Mike Fedyk
2003-07-16 22:08   ` 2.6.0-test1-mm1 William Lee Irwin III
2003-07-16 22:17     ` 2.6.0-test1-mm1 Mike Fedyk
2003-07-16 22:36   ` 2.6.0-test1-mm1 William Lee Irwin III
2003-07-16 22:56     ` 2.6.0-test1-mm1 Mike Fedyk
2003-07-17 11:57 2.6.0-test1-mm1 Downing, Thomas
2003-07-20 18:50 2.6.0-test1-mm1 Downing, Thomas

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:

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

  git send-email \ \ \ \ \ \ \ \

* 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).