From: Nick Piggin <nickpiggin@yahoo.com.au>
To: Paul Jackson <pj@engr.sgi.com>
Cc: Ingo Molnar <mingo@elte.hu>,
kenneth.w.chen@intel.com, torvalds@osdl.org, akpm@osdl.org,
linux-kernel@vger.kernel.org
Subject: Re: [patch] sched: auto-tune migration costs [was: Re: Industry db benchmark result on recent 2.6 kernels]
Date: Mon, 04 Apr 2005 12:08:21 +1000 [thread overview]
Message-ID: <4250A195.5030306@yahoo.com.au> (raw)
In-Reply-To: <20050403160807.35381385.pj@engr.sgi.com>
Paul Jackson wrote:
> Ingo wrote:
>
>>if you create a sched-domains hierarchy (based on the SLIT tables, or in
>>whatever other way) that matches the CPU hierarchy then you'll
>>automatically get the proper distances detected.
>
>
> Yes - agreed. I should push in the direction of improving the
> SN2 sched domain hierarchy.
>
I'd just be a bit careful about this. Your biggest systems will have
what? At least 7 or 8 domains if you're just going by the number of
hops, right? And maybe more if there is more to your topology than
just number of hops.
sched-domains firstly has a few problems even with your 2 level NUMA
domains (although I'm looking at fixing them if possible), but also
everything just has to do more work as you traverse the domains and
scan all CPUs for balancing opportunities. And its not like the cpu
scheduler uses any sort of exact science to make choices...
Nick
next prev parent reply other threads:[~2005-04-04 2:10 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-04-02 1:00 Industry db benchmark result on recent 2.6 kernels Chen, Kenneth W
2005-04-02 2:12 ` Nick Piggin
2005-04-02 14:53 ` [patch] sched: auto-tune migration costs [was: Re: Industry db benchmark result on recent 2.6 kernels] Ingo Molnar
2005-04-02 21:22 ` Paul Jackson
2005-04-03 5:53 ` Paul Jackson
2005-04-03 7:04 ` Ingo Molnar
2005-04-03 8:15 ` Paul Jackson
2005-04-03 11:34 ` Paul Jackson
2005-04-03 14:12 ` Paul Jackson
2005-04-03 15:01 ` Ingo Molnar
2005-04-03 22:30 ` Paul Jackson
2005-04-05 6:53 ` [patch] sched: auto-tune migration costs Andi Kleen
2005-04-05 7:20 ` Paul Jackson
2005-04-03 15:24 ` [patch] sched: auto-tune migration costs [was: Re: Industry db benchmark result on recent 2.6 kernels] Ingo Molnar
2005-04-03 23:08 ` Paul Jackson
2005-04-04 2:08 ` Nick Piggin [this message]
2005-04-04 3:55 ` Paul Jackson
2005-04-04 5:45 ` Ingo Molnar
2005-04-04 5:50 ` Paul Jackson
2005-04-04 5:56 ` Nick Piggin
2005-04-04 6:38 ` Paul Jackson
2005-04-04 6:48 ` Ingo Molnar
2005-04-04 7:37 ` Paul Jackson
2005-04-04 6:50 ` Ingo Molnar
2005-04-04 7:27 ` Paul Jackson
2005-04-03 14:29 ` Ingo Molnar
2005-04-03 23:15 ` Paul Jackson
2005-04-04 1:31 ` Chen, Kenneth W
2005-04-04 6:24 ` Ingo Molnar
2005-04-04 6:39 ` Ingo Molnar
2005-04-06 0:08 ` Chen, Kenneth W
2005-04-04 4:25 ` Andy Lutomirski
2005-04-04 4:36 ` Paul Jackson
2005-04-04 1:11 ` Chen, Kenneth W
2005-04-04 11:37 ` Ingo Molnar
2005-04-04 17:27 ` Paul Jackson
2005-04-05 1:43 ` Chen, Kenneth W
2005-04-05 1:49 ` Ingo Molnar
2005-04-05 3:04 ` Ingo Molnar
2005-04-06 3:33 ` Chen, Kenneth W
2005-04-06 6:45 ` Ingo Molnar
2005-04-08 2:27 ` Chen, Kenneth W
2005-04-03 9:01 ` Paul Jackson
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=4250A195.5030306@yahoo.com.au \
--to=nickpiggin@yahoo.com.au \
--cc=akpm@osdl.org \
--cc=kenneth.w.chen@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=pj@engr.sgi.com \
--cc=torvalds@osdl.org \
/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