From owner-freebsd-smp@FreeBSD.ORG Thu Mar 30 23:57:50 2006 Return-Path: X-Original-To: freebsd-smp@freebsd.org Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id B427516A401 for ; Thu, 30 Mar 2006 23:57:50 +0000 (UTC) (envelope-from luke@foolishgames.com) Received: from mail.foolishgames.com (mail.foolishgames.com [206.222.28.162]) by mx1.FreeBSD.org (Postfix) with ESMTP id 83C1D43D70 for ; Thu, 30 Mar 2006 23:57:45 +0000 (GMT) (envelope-from luke@foolishgames.com) Received: from [192.168.0.49] (24-247-121-206.dhcp.klmz.mi.charter.com [24.247.121.206]) (authenticated bits=0) by mail.foolishgames.com (8.13.6/8.13.3) with ESMTP id k2UNvg9V076057 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NO); Thu, 30 Mar 2006 18:57:43 -0500 (EST) (envelope-from luke@foolishgames.com) X-Habeas-Swe-9: mark in spam to . X-Habeas-Swe-6: email in exchange for a license for this Habeas Content-Transfer-Encoding: 7bit X-Habeas-Swe-3: like Habeas SWE (tm) Date: Thu, 30 Mar 2006 19:00:38 -0500 X-Priority: 3 X-Habeas-Swe-8: Message (HCM) and not spam. Please report use of this X-Habeas-Swe-5: Sender Warranted Email (SWE) (tm). The sender of this From: Lucas Holt X-Habeas-Swe-2: brightly anticipated In-Reply-To: <009301c65418$cc029f30$65fd24c0@Eric> References: <009301c65418$cc029f30$65fd24c0@Eric> To: Eric X-Habeas-Swe-7: warrant mark warrants that this is a Habeas Compliant Mime-Version: 1.0 (Apple Message framework v746.3) X-Habeas-Swe-4: Copyright 2002 Habeas (tm) Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed X-Habeas-Swe-1: winter into spring Message-Id: <64659004-E643-4044-AAED-076E0CC91977@foolishgames.com> X-Mailer: Apple Mail (2.746.3) X-Virus-Scanned: ClamAV 0.88/1364/Thu Mar 30 15:05:50 2006 on mail.foolishgames.com X-Virus-Status: Clean Cc: freebsd-smp@freebsd.org Subject: Re: Low perf with smp X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 30 Mar 2006 23:57:50 -0000 As i recall, mysql under FreeBSD 4.x used a userland threads library. That means its only capable of being scheduled on 1 cpu at a time and each thread is part of one process. I don't think SMP will help you much in this case. There have been a number of threads on some of the lists (hacker@ ?) in the last few months about MySQL performance indicating it might have something to do with the timer, etc. It might be worth looking into. Some people have stated that FreeBSD 5 and 6 are slower with disk io, etc. Plus you've got a raid 1 mirror which will slow down writes. 6.1 also has a reported issue with the 4bsd scheduler performance. Maybe thats part of the problem? On Mar 30, 2006, at 11:41 AM, Eric wrote: > Mail NETvigieHi all, > > I buy a new DELL PowerEdge 1850 with Bi-Xeon 2.8GHZ/2*2MB Dual Core > 800FSB processor, 2Go DDR2 RAM, PERC4di RAID controller with 2 SCSI > U320 15.000tpm 36Go hdd on RAID1. > > I install on it Freebsd 6.0, apache 2.2.0 and mysql 4.1.18 and > compare performance with : > OLD PowerEDge with only simple Bi-Xeon 3.0 GHZ and freebsd 4.11. > > The test is simple mysql query done by php script. > > Result is that test is two time slower with freebsd 6.0 and dual- > core than with old server. > I try with and without smp and threaded kernel and results are same. > I came back to Generic kernel to see if it come from my kernel > configuration but result was same. > > I then though it come from server and try to upgrade an other one > and have same result. > I then think it come from 6.0 and try with 6.1 beta4 and see same > result > I then downgrad to 4.11 without smp and then find good perfs. > So I just add to 4.11 generic kernel : > options SMP # Symmetric MultiProcessor > Kernel > options APIC_IO # Symmetric (APIC) I/O > > And then perf go back down from 30secondes to 54 sec. > I search a lot on internet and try with this : sysctl > machdep.cpu_idle_hlt=0 > But result is same. > > When doing whith smp top it's just as mysql have to wait to have > cpu. During 10,20 sec it just go to 2,4% then grow. Without smp it > grow very quickly. > > Does anyone know what's go wrong and what do I have to do to have > same perf with smp than without? > > Thanks, > > Eric. > > > _______________________________________________ > freebsd-smp@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-smp > To unsubscribe, send any mail to "freebsd-smp-unsubscribe@freebsd.org" Lucas Holt Luke@FoolishGames.com ________________________________________________________ FoolishGames.com (Jewel Fan Site) JustJournal.com (Free blogging) FoolishGames.net (Enemy Territory site)