From: Peter Williams <pwil3058@bigpond.net.au>
To: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Andrew Morton <akpm@osdl.org>, Ingo Molnar <mingo@elte.hu>,
Nick Piggin <nickpiggin@yahoo.com.au>,
Con Kolivas <kernel@kolivas.org>,
"Chen, Kenneth W" <kenneth.w.chen@intel.com>,
Mike Galbraith <efault@gmx.de>,
linux-kernel@vger.kernel.org
Subject: Re: more smpnice patch issues
Date: Sat, 25 Mar 2006 11:56:27 +1100 [thread overview]
Message-ID: <4424953B.9030000@bigpond.net.au> (raw)
In-Reply-To: <20060324154558.A20018@unix-os.sc.intel.com>
Siddha, Suresh B wrote:
> more issues with smpnice patch...
>
> a) consider a 4-way system (simple SMP system with no HT and cores) scenario
> where a high priority task (nice -20) is running on P0 and two normal
> priority tasks running on P1. load balance with smp nice code
> will never be able to detect an imbalance and hence will never move one of
> the normal priority tasks on P1 to idle cpus P2 or P3.
Why?
>
> b) smpnice seems to break this patch..
>
> [PATCH] sched: allow the load to grow upto its cpu_power
> http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=0c117f1b4d14380baeed9c883f765ee023da8761
>
> example scenario for this case: consider a numa system with two nodes, each
> node containing four processors. if there are two processes in node-0 and with
> node-1 being completely idle, your patch will move one of those processes to
> node-1 whereas the previous behavior will retain those two processes in node-0..
> (in this case, in your code max_load will be less than busiest_load_per_task)
>
> smpnice patch has complicated the load balance code... Very difficult
> to comprehend the side effects of this patch in the presence of different
> priority tasks...
That is NOT true. Without smpnice whether the "right thing" happens
with non zero nice tasks is largely down to luck. With smpnice the
result is far more predictable.
The PURPOSE of smpnice IS to alter load balancing in the face of the use
of non zero nice tasks. The reason for doing this is so that nice
reliably effects the allocation of CPU resources on SMP machines. I.e.
changes in load balancing behaviour as a result of tasks having nice
values other than zero is the desired result and NOT a side effect.
Peter
--
Peter Williams pwil3058@bigpond.net.au
"Learning, n. The kind of ignorance distinguishing the studious."
-- Ambrose Bierce
next prev parent reply other threads:[~2006-03-25 0:56 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-03-22 23:51 cpu scheduler merge plans Andrew Morton
2006-03-22 22:57 ` kernel
2006-03-23 1:37 ` Siddha, Suresh B
2006-03-23 22:06 ` Peter Williams
2006-03-23 0:31 ` Nick Piggin
2006-03-23 1:16 ` Peter Williams
2006-03-24 23:45 ` more smpnice patch issues Siddha, Suresh B
2006-03-25 0:56 ` Peter Williams [this message]
2006-03-25 1:53 ` Peter Williams
2006-03-25 3:40 ` [PATCH] sched: make sure busiest group and run queue are pullable Peter Williams
2006-03-26 23:24 ` more smpnice patch issues Peter Williams
2006-03-26 23:43 ` [PATCH] sched: smpnice prevent integer arithmetic wrap problems Peter Williams
2006-03-23 5:03 ` cpu scheduler merge plans Ingo Molnar
2006-03-23 5:13 ` Andrew Morton
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=4424953B.9030000@bigpond.net.au \
--to=pwil3058@bigpond.net.au \
--cc=akpm@osdl.org \
--cc=efault@gmx.de \
--cc=kenneth.w.chen@intel.com \
--cc=kernel@kolivas.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=nickpiggin@yahoo.com.au \
--cc=suresh.b.siddha@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 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.