From nobody@FreeBSD.org  Wed Aug 27 10:51:25 2008
Return-Path: <nobody@FreeBSD.org>
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34])
	by hub.freebsd.org (Postfix) with ESMTP id AB74F106564A
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 27 Aug 2008 10:51:24 +0000 (UTC)
	(envelope-from nobody@FreeBSD.org)
Received: from www.freebsd.org (www.freebsd.org [IPv6:2001:4f8:fff6::21])
	by mx1.freebsd.org (Postfix) with ESMTP id A9BE98FC2B
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 27 Aug 2008 10:51:24 +0000 (UTC)
	(envelope-from nobody@FreeBSD.org)
Received: from www.freebsd.org (localhost [127.0.0.1])
	by www.freebsd.org (8.14.2/8.14.2) with ESMTP id m7RApOk8013822
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 27 Aug 2008 10:51:24 GMT
	(envelope-from nobody@www.freebsd.org)
Received: (from nobody@localhost)
	by www.freebsd.org (8.14.2/8.14.1/Submit) id m7RApODg013821;
	Wed, 27 Aug 2008 10:51:24 GMT
	(envelope-from nobody)
Message-Id: <200808271051.m7RApODg013821@www.freebsd.org>
Date: Wed, 27 Aug 2008 10:51:24 GMT
From: KOIE Hidetaka <koie@suri.co.jp>
To: freebsd-gnats-submit@FreeBSD.org
Subject: CPU usages are unbalanced
X-Send-Pr-Version: www-3.1
X-GNATS-Notify:

>Number:         126880
>Category:       kern
>Synopsis:       [scheduler] CPU usages are unbalanced
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          change-request
>Submitter-Id:   current-users
>Arrival-Date:   Wed Aug 27 11:00:05 UTC 2008
>Closed-Date:    Sun Feb 01 01:27:28 UTC 2009
>Last-Modified:  Sun Feb 01 01:27:28 UTC 2009
>Originator:     KOIE Hidetaka
>Release:        8.0-CURRENT
>Organization:
surigiken
>Environment:
FreeBSD guriandgura 8.0-CURRENT FreeBSD 8.0-CURRENT #0: Thu Aug 21 12:16:00 JST 2008     koie@guriandgura:/usr/obj/usr/src/sys/GURIANDGURA  amd64

>Description:
Idle times per CPU can be watched by top(1).
But these times are clearly different.

>How-To-Repeat:
This machine is dual-core opteron and dual-socket.
koie@guriandgura% top -HSn|grep idle
   11 root       171 ki31     0K    64K RUN    0  26.7H 41.55% {idle: cpu0}
   11 root       171 ki31     0K    64K RUN    2  23.5H 31.88% {idle: cpu2}
   11 root       171 ki31     0K    64K RUN    1  19.9H 15.19% {idle: cpu1}
   11 root       171 ki31     0K    64K RUN    3  20.8H  9.18% {idle: cpu3}
koie@guriandgura%

>Fix:


>Release-Note:
>Audit-Trail:

From: "Remko Lodder" <remko@elvandar.org>
To: "KOIE Hidetaka" <koie@suri.co.jp>
Cc: freebsd-gnats-submit@freebsd.org
Subject: Re: kern/126880: CPU usages are unbalanced
Date: Wed, 27 Aug 2008 13:31:36 +0200 (CEST)

 On Wed, August 27, 2008 12:51 pm, KOIE Hidetaka wrote:
 >
 >>How-To-Repeat:
 > This machine is dual-core opteron and dual-socket.
 > koie@guriandgura% top -HSn|grep idle
 >    11 root       171 ki31     0K    64K RUN    0  26.7H 41.55% {idle:
 > cpu0}
 >    11 root       171 ki31     0K    64K RUN    2  23.5H 31.88% {idle:
 > cpu2}
 >    11 root       171 ki31     0K    64K RUN    1  19.9H 15.19% {idle:
 > cpu1}
 >    11 root       171 ki31     0K    64K RUN    3  20.8H  9.18% {idle:
 > cpu3}
 > koie@guriandgura%
 >
 
 Hello,
 
 What exactly do you want to show/tell us with this? Yes the  idle times
 are unbalanced, some CPU's can be more busy then others, and thus have
 less IDLE time recorded and less CPU spending on being idle (sounds
 contradictive :-)), but I do not see that as a problem, remember that
 there are unthreaded applications that just work on a single CPU , which
 might have been CPU3....
 
 Thanks,
 remko
 
 -- 
 /"\   Best regards,                      | remko@FreeBSD.org
 \ /   Remko Lodder                       | remko@EFnet
  X    http://www.evilcoder.org/          |
 / \   ASCII Ribbon Campaign              | Against HTML Mail and News
 
 

From: Kris Kennaway <kris@FreeBSD.org>
To: KOIE Hidetaka <koie@suri.co.jp>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/126880: CPU usages are unbalanced
Date: Wed, 27 Aug 2008 16:14:15 +0200

 KOIE Hidetaka wrote:
 > Idle times per CPU can be watched by top(1).
 > But these times are clearly different.
 > 
 >> How-To-Repeat:
 > This machine is dual-core opteron and dual-socket.
 > koie@guriandgura% top -HSn|grep idle
 >    11 root       171 ki31     0K    64K RUN    0  26.7H 41.55% {idle: cpu0}
 >    11 root       171 ki31     0K    64K RUN    2  23.5H 31.88% {idle: cpu2}
 >    11 root       171 ki31     0K    64K RUN    1  19.9H 15.19% {idle: cpu1}
 >    11 root       171 ki31     0K    64K RUN    3  20.8H  9.18% {idle: cpu3}
 
 Do you have evidence that this is causing a performance problem?  Your 
 system is not 100% busy anyway, and it is often better to keep a 
 long-running task running on the same CPU where it can benefit from hot 
 caches rather than bouncing it around between CPUs where it will suffer 
 from lots of cache misses.
 
 Kris
 
State-Changed-From-To: open->closed 
State-Changed-By: linimon 
State-Changed-When: Sun Feb 1 01:26:47 UTC 2009 
State-Changed-Why:  
apparently this is working as designed. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=126880 
>Unformatted:
