From owner-freebsd-performance@FreeBSD.ORG Sun Jan 2 06:01:54 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7FAE016A4CE for ; Sun, 2 Jan 2005 06:01:54 +0000 (GMT) Received: from webmail.uoi.gr (webmail.uoi.gr [195.130.120.121]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2704343D39 for ; Sun, 2 Jan 2005 06:01:53 +0000 (GMT) (envelope-from dkouroun@cc.uoi.gr) Received: from webmail.uoi.gr (localhost [127.0.0.1])j0261pNP025480 for ; Sun, 2 Jan 2005 08:01:51 +0200 Received: (from wwwrun@localhost) by webmail.uoi.gr (8.12.10/8.12.10/Submit) id j0261pVo025479 for freebsd-performance@freebsd.org; Sun, 2 Jan 2005 08:01:51 +0200 Received: from 212-70-197-171.rod.dialup.tee.gr (212-70-197-171.rod.dialup.tee.gr [212.70.197.171]) by webmail.uoi.gr (IMP) with HTTP for ; Sun, 2 Jan 2005 08:01:51 +0200 Message-ID: <1104645711.41d78e4f22258@webmail.uoi.gr> Date: Sun, 2 Jan 2005 08:01:51 +0200 From: dkouroun@cc.uoi.gr To: freebsd-performance@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.2 X-Originating-IP: 212.70.197.171 Subject: nvidia-drivers with 5.3 and 4.10 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 02 Jan 2005 06:01:54 -0000 Hi all! I am writing a program in OpenGL both in SuSE 9.0 with the latest nvidia drivers NVIDIA-Linux-x86-1.0-6629-pkg1.run and in FreeBSD 5.3, 4.10 with NVIDIA-FreeBSD-x86-1.0-6113.tar.gz drivers. I used in both systems the option Option "NvAGP" "2". The openGL program was compiled in both systems with gcc-3.4.3 (-O2 -fomit-frame-pointer) The results are: Linux SuSE 9.0 : 330 Frames Per Second FreeBSD 5.3 Release : 92 FreeBSD 4.10 : 100 Any ideas? I also reinstalled Nvidia drivers in SuSE just to make sure that it is not the driver which makes the difference so I downloaded and installed a version much older than one I used above: NVIDIA-Linux-x86-1.0-6111-pkg1.run The results in performance were the same for Linux! What is wrong in FreeBSD? Best Wishes to Everybody for the new year! D.K. From owner-freebsd-performance@FreeBSD.ORG Sun Jan 2 11:08:26 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C55B316A4CE for ; Sun, 2 Jan 2005 11:08:26 +0000 (GMT) Received: from webmail.uoi.gr (webmail.uoi.gr [195.130.120.121]) by mx1.FreeBSD.org (Postfix) with ESMTP id BA27743D53 for ; Sun, 2 Jan 2005 11:08:24 +0000 (GMT) (envelope-from dkouroun@cc.uoi.gr) Received: from webmail.uoi.gr (localhost [127.0.0.1])j02B8LNP027188; Sun, 2 Jan 2005 13:08:22 +0200 Received: (from wwwrun@localhost) by webmail.uoi.gr (8.12.10/8.12.10/Submit) id j02B8LlC027187; Sun, 2 Jan 2005 13:08:21 +0200 Received: from 212-70-197-171.rod.dialup.tee.gr (212-70-197-171.rod.dialup.tee.gr [212.70.197.171]) by webmail.uoi.gr (IMP) with HTTP for ; Sun, 2 Jan 2005 13:08:21 +0200 Message-ID: <1104664101.41d7d6257f3dd@webmail.uoi.gr> Date: Sun, 2 Jan 2005 13:08:21 +0200 From: dkouroun@cc.uoi.gr To: freebsd-performance@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.2 X-Originating-IP: 212.70.197.171 Subject: nvidia-drivers with 4.10 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 02 Jan 2005 11:08:26 -0000 Dear freebsd-performance people, I went on measuring performance in 4.10 Release. I compiled the lates FreeBSD drivers by nvidia. The problem was that with FreeBSD 4.10 native agp the performance was unacceptable 5 frames per second. I rebooted and changed /etc/X11/XF86Config such that it uses nvidia agp (Option "NvAGP" "1"). The first time I run the program the performance went up to 117fps. Have in mind that in SuSE it is 269fps. The second time I run the program the performance dropped down to 5fps again! Can anybody explain what is going on? Thanks in advance! Drosos. From owner-freebsd-performance@FreeBSD.ORG Mon Jan 3 19:33:31 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D1F4516A4CE for ; Mon, 3 Jan 2005 19:33:30 +0000 (GMT) Received: from stewie.obfuscated.net (stewie.obfuscated.net [66.118.188.125]) by mx1.FreeBSD.org (Postfix) with ESMTP id 92F1A43D2D for ; Mon, 3 Jan 2005 19:33:30 +0000 (GMT) (envelope-from meconlen@obfuscated.net) Received: from [192.168.1.101] (653259hfc120.tampabay.rr.com [65.32.59.120]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by stewie.obfuscated.net (Postfix) with ESMTP id EA75B6137 for ; Mon, 3 Jan 2005 14:33:29 -0500 (EST) Mime-Version: 1.0 (Apple Message framework v619) Content-Transfer-Encoding: 7bit Message-Id: <3F362419-5DBE-11D9-B88F-00039367611E@obfuscated.net> Content-Type: text/plain; charset=US-ASCII; format=flowed To: freebsd-performance@freebsd.org From: Michael E.Conlen Date: Mon, 3 Jan 2005 14:32:46 -0500 X-Mailer: Apple Mail (2.619) Subject: FC Disk array X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 03 Jan 2005 19:33:31 -0000 I have a system connected to a IBM Fast-t 100 fiber channel disk array that I'm using as an NFS server. It's handling a moderate load of write operations, normally in the range of 4 to 8 MB/sec with an average write size of 16k constantly. The odd thing I've noticed in gstat and systat is that the system reports the disk as over 100% busy at times. I once saw it as high as 150%. I don't care that it does this, but it makes me wonder what the ceiling is. The system is a dual Xeon HT with hlt_logical_cpu set to 0 so I'm used to getting 400% worth of CPU in some of my metrics. -- Michael Conlen meconlen@obfuscated.net From owner-freebsd-performance@FreeBSD.ORG Tue Jan 4 13:33:49 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 418E716A4CF for ; Tue, 4 Jan 2005 13:33:49 +0000 (GMT) Received: from rproxy.gmail.com (rproxy.gmail.com [64.233.170.206]) by mx1.FreeBSD.org (Postfix) with ESMTP id BED0F43D39 for ; Tue, 4 Jan 2005 13:33:48 +0000 (GMT) (envelope-from molnarcs@gmail.com) Received: by rproxy.gmail.com with SMTP id g11so311998rne for ; Tue, 04 Jan 2005 05:33:48 -0800 (PST) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=Dea/E7ZMEoOfYQjjeaugQXABvGy8nahwqCERpQpciVW8UKnhUUmbCdsdTA/sIMHUUpac8x0o6iGwbk/lLkHjFG1bSoTztI43NqZYwushHrBpN+C6OSZMG8V+AMOWoKIkOMF9T/Y49Om4tsfbnsWphwBoeFib8a2+kW3Jz7trDoo= Received: by 10.38.207.31 with SMTP id e31mr734535rng; Tue, 04 Jan 2005 05:33:48 -0800 (PST) Received: by 10.38.207.37 with HTTP; Tue, 4 Jan 2005 05:33:48 -0800 (PST) Message-ID: Date: Tue, 4 Jan 2005 14:33:48 +0100 From: =?ISO-8859-1?Q?Csaba_Moln=E1r?= To: dkouroun@cc.uoi.gr, freebsd-performance@freebsd.org Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Mailman-Approved-At: Tue, 04 Jan 2005 15:01:19 +0000 Subject: nvidia on FreeBSD 5.3 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: =?ISO-8859-1?Q?Csaba_Moln=E1r?= List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 Jan 2005 13:33:49 -0000 I'm not on the list, that's why I mail directly... I noticed some strange behaviour on FreeBSD 5.3. I had these settings in my /etc/sysctl.conf hw.nvidia.registry.EnableVia4x=1 hw.nvidia.registry.EnableAGPSBA=1 hw.nvidia.registry.EnableAGPFW=1 However, EnableAGPSBA/AGPFW was still not set after boot up for some strange reason. glxgears with this setting reported ~250-300 fps. Card is: GeForce FX 5200/128 (64 bit). After changing registry values manually, (sysctl hw.nvidia.registry.EnableVia4x=1 ...) fps is now: mcsaba@mcsaba$ glxgears 3319 frames in 5.0 seconds = 663.800 FPS 4537 frames in 5.0 seconds = 907.400 FPS 4528 frames in 5.0 seconds = 905.600 FPS 4506 frames in 5.0 seconds = 901.200 FPS 4544 frames in 5.0 seconds = 908.800 FPS X connection to :0.0 broken (explicit kill or server shutdown). mcsaba@mcsaba$ I put hw.nvidia.registry.EnableAGPSBA=1 hw.nvidia.registry.EnableAGPFW=1 in /boot/loader.conf, and now they seem to be set at boot up, however, sysctl -a | grep nvidia still reports status as disabled. I know it isn't because there is a substantial difference, but here is my output anyway: root@mcsaba# sysctl -a | grep nvidia nvidia 348 738K 791K 6419 16,32,64,128,256,512,1024,2048,4096 hw.nvidia.agp.card.rates: 4x 2x 1x hw.nvidia.agp.card.fw: supported hw.nvidia.agp.card.sba: supported hw.nvidia.agp.card.registers: 0x1f000217:0x1f000104 hw.nvidia.agp.status.status: enabled hw.nvidia.agp.status.driver: freebsd (agp.ko) hw.nvidia.agp.status.rate: 4x hw.nvidia.agp.status.fw: disabled hw.nvidia.agp.status.sba: disabled hw.nvidia.version: NVIDIA FreeBSD x86 NVIDIA Kernel Module 1.0-6113 Mon Aug 2 16:08:32 PDT 2004 hw.nvidia.registry.EnableVia4x: 1 hw.nvidia.registry.EnableALiAGP: 0 hw.nvidia.registry.NvAGP: 3 hw.nvidia.registry.EnableAGPSBA: 1 hw.nvidia.registry.EnableAGPFW: 1 hw.nvidia.registry.SoftEDIDs: 1 hw.nvidia.registry.Mobile: 4294967295 hw.nvidia.registry.ResmanDebugLevel: 4294967295 hw.nvidia.registry.FlatPanelMode: 0 hw.nvidia.cards.0.model: GeForce FX 5200 hw.nvidia.cards.0.irq: 16 hw.nvidia.cards.0.vbios: 04.34.20.27.06 hw.nvidia.cards.0.type: AGP dev.nvidia.0.%desc: GeForce FX 5200 dev.nvidia.0.%driver: nvidia dev.nvidia.0.%location: slot=0 function=0 dev.nvidia.0.%pnpinfo: vendor=0x10de device=0x0322 subvendor=0x0000 subdevice=0x0000 class=0x030000 dev.nvidia.0.%parent: pci1 I compiled the nvidia-driver port WITH_ACPI=yes WITH_FREEBSD_AGP=yes WITH_LINUX=yes I don't know what might be the problem in your setting. I just thought that it might help if you compare it with my system. Where did you measure those fps? If it was with your own prog, than it must be something with your own prog. (you might try glxgears on both linux and freebsd as well). Unfortunately, I'm not a programmer, so I can't help much with that. Maybe if you post more details someone else might be more helpful :) From owner-freebsd-performance@FreeBSD.ORG Tue Jan 4 18:58:02 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3C73616A4CE for ; Tue, 4 Jan 2005 18:58:02 +0000 (GMT) Received: from sb.santaba.com (sb.santaba.com [207.154.84.26]) by mx1.FreeBSD.org (Postfix) with ESMTP id 13DBD43D39 for ; Tue, 4 Jan 2005 18:58:02 +0000 (GMT) (envelope-from jbehl@fastclick.com) Received: from [192.168.3.100] (unknown [205.180.85.193]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by sb.santaba.com (Postfix) with ESMTP id BE3DA28433; Tue, 4 Jan 2005 10:58:01 -0800 (PST) Message-ID: <41DAE753.7050800@fastclick.com> Date: Tue, 04 Jan 2005 10:58:27 -0800 From: Jeff Behl User-Agent: Mozilla Thunderbird 0.8 (Windows/20040913) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Sean Chittenden References: <7632915A8F000C4FAEFCF272A880344165164F@Ehost067.exch005intermedia.net> <298151B6-5527-11D9-B830-000A95C705DC@chittenden.org> In-Reply-To: <298151B6-5527-11D9-B830-000A95C705DC@chittenden.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: %cpu in system - squid performance in FreeBSD 5.3 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 Jan 2005 18:58:02 -0000 back from vacation... i did try STABLE, with the same results: FreeBSD www2.cdn 5.3-STABLE FreeBSD 5.3-STABLE #2: Sun Dec 5 21:06:14 PST 2004 root@www2.cdn:/usr/obj/usr/src/sys/SMP amd64 interestingly enough, i got a reply on a squid list from someone on a linux box who had similar results: - I've got a dual proc Xeon (3GHz) Linux 2.4.26 system running one squid 2.5.5 as reverse proxy without cache, I use squidguard to route different path of the same domain to different clusters. Top shows: Cpu1 : 39.3% user, 58.0% system, 0.0% nice, 2.7% idle - so perhaps this is some gross problem with squid? jeff Sean Chittenden wrote: >> As a follow up to the below (original message at the very bottom), I >> installed a load balancer in front of the machines which terminates the >> tcp connections from clients and opens up a few, persistent connections >> to each server over which requests are pipelined. In this scenario >> everything is copasetic: > > > Interesting... I wonder what the lock contention path is between the > VM and network stack. Has anyone else noticed this in their testing? > Did you try a post-5.3 release or not? rwatson@ just MFC'ed a bunch > of network locking fixes to RELENG_5, but none-stand out in my mind as > being something that'd potentially fix your issue. Actually, he > probably knows better than anyone what this could be. Some of the > post RELENG_5_3 commits by alc@ could explain this, however, and is > why I ask what the specific version/build time is for your release. -sc > From owner-freebsd-performance@FreeBSD.ORG Tue Jan 4 19:19:11 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0362416A4CE; Tue, 4 Jan 2005 19:19:11 +0000 (GMT) Received: from mail.trippynames.com (mail.trippynames.com [38.113.223.19]) by mx1.FreeBSD.org (Postfix) with ESMTP id A59A643D55; Tue, 4 Jan 2005 19:19:10 +0000 (GMT) (envelope-from sean@chittenden.org) Received: from localhost (localhost [127.0.0.1]) by mail.trippynames.com (Postfix) with ESMTP id C4CAAA8121; Tue, 4 Jan 2005 11:18:04 -0800 (PST) Received: from mail.trippynames.com ([127.0.0.1]) by localhost (rand.nxad.com [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 33568-01; Tue, 4 Jan 2005 11:17:47 -0800 (PST) Received: from [192.168.1.5] (dsl081-069-073.sfo1.dsl.speakeasy.net [64.81.69.73]) by mail.trippynames.com (Postfix) with ESMTP id B0480A81B2; Tue, 4 Jan 2005 11:17:27 -0800 (PST) In-Reply-To: <7632915A8F000C4FAEFCF272A880344165164F@Ehost067.exch005intermedia.net> References: <7632915A8F000C4FAEFCF272A880344165164F@Ehost067.exch005intermedia.net> Mime-Version: 1.0 (Apple Message framework v619) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <69A7A6D4-5E85-11D9-ACFF-000A95C705DC@chittenden.org> Content-Transfer-Encoding: 7bit From: Sean Chittenden Date: Tue, 4 Jan 2005 11:18:28 -0800 To: "Jeff Behl" X-Mailer: Apple Mail (2.619) cc: freebsd-net@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: %cpu in system - squid performance in FreeBSD 5.3 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 Jan 2005 19:19:11 -0000 > PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU > COMMAND > 474 squid 96 0 68276K 62480K select 0 53:38 16.80% 16.80% > squid > 311 bind 20 0 10628K 6016K kserel 0 12:28 0.00% 0.00% > named > > It's actually so good that one machine can now handle all traffic > (around 180 Mb/s) at < %50 cpu utilization. Seems like something in > the > network stack is responsible for the high %system cpu util... OH!!!! Wow, I should've noticed that earlier. Your hint about someone on Linux having the same problem tipped me off to look at the process state again. Anyway, you nearly answered your own question, save you probably aren't familiar with select(2)'s lack of scalability. Read these: http://www.kegel.com/c10k.html Specifically the four methods mentioned here: http://www.kegel.com/c10k.html#nb.select Then look at the benchmarks done using libevent(3): http://www.monkey.org/~provos/libevent/ Dime to dollar you're spending all of your time copying file descriptor arrays in and out of the kernel because squid uses select(2) instead of kqueue(2). Might be an interesting project for someone to take up to convert that to kqueue(2). Until then, any local TCP load balancer that uses kqueue(2) would also solve your problem (I'm not aware of any off the top of my head... pound(8) does, but it is only used for HTTP and is not a reverse proxy) and would likely prevent you from having your problems. -sc -- Sean Chittenden From owner-freebsd-performance@FreeBSD.ORG Tue Jan 4 22:22:31 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 80C7916A4CE for ; Tue, 4 Jan 2005 22:22:31 +0000 (GMT) Received: from mail2.dbitech.ca (radius.wavefire.com [64.141.13.252]) by mx1.FreeBSD.org (Postfix) with SMTP id D865E43D2D for ; Tue, 4 Jan 2005 22:22:30 +0000 (GMT) (envelope-from darcy@wavefire.com) Received: (qmail 10522 invoked from network); 4 Jan 2005 23:35:47 -0000 Received: from dbitech.wavefire.com (HELO ?64.141.15.253?) (darcy@64.141.15.253) by radius.wavefire.com with SMTP; 4 Jan 2005 23:35:47 -0000 From: Darcy Buskermolen Organization: Wavefire Technologies Corp. To: freebsd-net@freebsd.org Date: Tue, 4 Jan 2005 14:22:29 -0800 User-Agent: KMail/1.6.2 References: <7632915A8F000C4FAEFCF272A880344165164F@Ehost067.exch005intermedia.net> <69A7A6D4-5E85-11D9-ACFF-000A95C705DC@chittenden.org> In-Reply-To: <69A7A6D4-5E85-11D9-ACFF-000A95C705DC@chittenden.org> MIME-Version: 1.0 Content-Disposition: inline Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <200501041422.29131.darcy@wavefire.com> cc: Jeff Behl cc: freebsd-performance@freebsd.org Subject: Re: %cpu in system - squid performance in FreeBSD 5.3 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 Jan 2005 22:22:31 -0000 On January 4, 2005 11:18 am, Sean Chittenden wrote: > > PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU > > COMMAND > > 474 squid 96 0 68276K 62480K select 0 53:38 16.80% 16.80% > > squid > > 311 bind 20 0 10628K 6016K kserel 0 12:28 0.00% 0.00% > > named > > > > It's actually so good that one machine can now handle all traffic > > (around 180 Mb/s) at < %50 cpu utilization. Seems like something in > > the > > network stack is responsible for the high %system cpu util... > > OH!!!! Wow, I should've noticed that earlier. Your hint about someone > on Linux having the same problem tipped me off to look at the process > state again. Anyway, you nearly answered your own question, save you > probably aren't familiar with select(2)'s lack of scalability. Read > these: > > http://www.kegel.com/c10k.html > > > Specifically the four methods mentioned here: > > http://www.kegel.com/c10k.html#nb.select > > > Then look at the benchmarks done using libevent(3): > > http://www.monkey.org/~provos/libevent/ > > > Dime to dollar you're spending all of your time copying file descriptor > arrays in and out of the kernel because squid uses select(2) instead of > kqueue(2). Might be an interesting project for someone to take up to > convert that to kqueue(2). Until then, any local TCP load balancer > that uses kqueue(2) would also solve your problem (I'm not aware of any > off the top of my head... pound(8) does, but it is only used for HTTP > and is not a reverse proxy) and would likely prevent you from having > your problems. -sc squid-dev has kqueue support already. ./configure --disable-select --disable-poll --enable-kqueue -- Darcy Buskermolen Wavefire Technologies Corp. ph: 250.717.0200 fx: 250.763.1759 http://www.wavefire.com From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 01:19:46 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E822B16A4CE for ; Thu, 6 Jan 2005 01:19:46 +0000 (GMT) Received: from rfhs0023.fh-regensburg.de (rfhs0023.fh-regensburg.de [194.95.104.23]) by mx1.FreeBSD.org (Postfix) with SMTP id BFC5243D2D for ; Thu, 6 Jan 2005 01:19:45 +0000 (GMT) (envelope-from hubert@feyrer.de) Received: (qmail 20838 invoked from network); 6 Jan 2005 01:19:41 -0000 Received: from unknown (HELO rfhs0023.fh-regensburg.de) (127.0.0.1) by 0 with SMTP; 6 Jan 2005 01:19:41 -0000 Received: from rfhpc8323.fh-regensburg.de (rfhpc8323.fh-regensburg.de [194.95.108.191])j061JckY020833; Thu, 6 Jan 2005 02:19:39 +0100 (MET) (envelope-from hubert@feyrer.de) Received: from rfhpc8317 (rfhpc8317.fh-regensburg.de [194.95.108.65]) j061Jca21924; Thu, 6 Jan 2005 02:19:38 +0100 (MET) Date: Thu, 6 Jan 2005 02:19:37 +0100 (CET) From: Hubert Feyrer X-X-Sender: feyrer@rfhpc8317 To: freebsd-performance@FreeBSD.org, freebsd-advocacy@FreeBSD.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-PMX-Version: 4.7.0.111621, Antispam-Engine: 2.0.2.0, Antispam-Data: 2005.1.5.26 X-PerlMx-Spam: Gauge=IIIIIII, Probability=7%, Report='__CT 0, __CT_TEXT_PLAIN 0, __HAS_MSGID 0, __MIME_VERSION 0, __SANE_MSGID 0' X-Mailman-Approved-At: Thu, 06 Jan 2005 02:24:56 +0000 Subject: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 01:19:47 -0000 Gregory McGarry asked me to host and advertize this paper for him: Abstract: ``With the recent releases of NetBSD 2.0 and FreeBSD 5.3 operating system, many new and exciting features have been implemented. Both criticism and commendation on performance, reliability and scalability have been directed towards these releases. This paper presents a suite of benchmarks and results for comparing the performance of these operating systems. The benchmarks target core operating system functionality, server scalability and thread implementation. These benchmarks are useful server-based criteria for demanding applications such as loaded webservers, databases, and voice-over-IP (VoIP) media relays. The results indicate that NetBSD has surpassed FreeBSD in performance on nearly every benchmark and is poised to grab the title of the best operating system for the server environment.'' URL: http://www.feyrer.de/NetBSD/gmcgarry/ - Hubert -- NetBSD - Free AND Open! (And of course secure, portable, yadda yadda) From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 02:52:26 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0ED5A16A4CE for ; Thu, 6 Jan 2005 02:52:26 +0000 (GMT) Received: from f7.mail.ru (f7.mail.ru [194.67.57.37]) by mx1.FreeBSD.org (Postfix) with ESMTP id A15C743D1D for ; Thu, 6 Jan 2005 02:52:25 +0000 (GMT) (envelope-from shmukler@mail.ru) Received: from mail by f7.mail.ru with local id 1CmNlA-0008C3-00; Thu, 06 Jan 2005 05:52:24 +0300 Received: from [24.185.78.105] by mac.mail.ru with HTTP; Thu, 06 Jan 2005 05:52:24 +0300 From: Igor Shmukler To: Hubert Feyrer Mime-Version: 1.0 X-Mailer: mPOP Web-Mail 2.19 X-Originating-IP: [24.185.78.105] Date: Thu, 06 Jan 2005 05:52:24 +0300 In-Reply-To: Content-Type: text/plain; charset=koi8-r Content-Transfer-Encoding: 8bit Message-Id: cc: freebsd-performance@FreeBSD.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Igor Shmukler List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 02:52:26 -0000 Hubert, The benchmark suite in question is not bad, but it's a microbenchmark suite. BTW, when we used the suite to tune kernel for ourt needs we had to change many things. As of a year ago some tests had to be modified, otherwise we might start tuning for the wrong goals. The title of this email and the article is NetBSD beats FreeBSD in performance should be changed to NetBSD beats FreeBSD on comprehensive microbenchmark suite. I don't remember who it was, but someone posted results of real server tests. These were in favor of FreeBSD. I doubt that much changed since last year but who knows ... Maybe NetBSD is better, but this is not something that could be concluded from a microbenchmark. IMHO. igor. -----Original Message----- From: Hubert Feyrer To: freebsd-performance@FreeBSD.org, freebsd-advocacy@FreeBSD.org Date: Thu, 6 Jan 2005 02:19:37 +0100 (CET) Subject: Benchmark: NetBSD 2.0 be ats FreeBSD 5.3 in server performance > > > Gregory McGarry asked me to host and advertize this paper for him: > > Abstract: ``With the recent releases of NetBSD 2.0 and FreeBSD 5.3 operating > system, many new and exciting features have been implemented. Both criticism > and commendation on performance, reliability and scalability have been directed > towards these releases. > > This paper presents a suite of benchmarks and results for comparing the > performance of these operating systems. The benchmarks target core operating > system functionality, server scalability and thread implementation. These > benchmarks are useful server-based criteria for demanding applications such as > loaded webservers, databases, and voice-over-IP (VoIP) media relays. The > results indicate that NetBSD has surpassed FreeBSD in performance on nearly > every benchmark and is poised to grab the title of the best o perating system > for the server environment.'' > > URL: http://www.feyrer.de/NetBSD/gmcgarry/ > > > - Hubert > > -- > NetBSD - Free AND Open! (And of course secure, portable, yadda yadda) > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > http://Mail.Ru - лучшая почта с неограниченным объемом почтового ящика! From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 03:18:53 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A442716A4CE for ; Thu, 6 Jan 2005 03:18:53 +0000 (GMT) Received: from stephanie.unixdaemons.com (stephanie.unixdaemons.com [67.18.111.194]) by mx1.FreeBSD.org (Postfix) with ESMTP id 380DD43D48 for ; Thu, 6 Jan 2005 03:18:53 +0000 (GMT) (envelope-from bmilekic@technokratis.com) Received: from stephanie.unixdaemons.com (bmilekic@localhost.unixdaemons.com [127.0.0.1])j063IlTF022732; Wed, 5 Jan 2005 22:18:47 -0500 (EST) Received: (from bmilekic@localhost) by stephanie.unixdaemons.com (8.13.2/8.12.1/Submit) id j063IlKX022730; Wed, 5 Jan 2005 22:18:47 -0500 (EST) (envelope-from bmilekic@technokratis.com) X-Authentication-Warning: stephanie.unixdaemons.com: bmilekic set sender to bmilekic@technokratis.com using -f Date: Wed, 5 Jan 2005 22:18:47 -0500 From: Bosko Milekic To: Igor Shmukler Message-ID: <20050106031847.GA22306@technokratis.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: User-Agent: Mutt/1.4.2.1i cc: Hubert Feyrer cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 03:18:53 -0000 Not to mention, FreeBSD5 has yet to be micro-optimized. How about some scalability benchmarks on multi-CPU machines? The original post (particular since it was sent to -advocacy) is FUD. -Bosko On Thu, Jan 06, 2005 at 05:52:24AM +0300, Igor Shmukler wrote: > Hubert, The benchmark suite in question is not bad, but it's a microb= enchmark suite. BTW, when we used the suite to tune kernel for ourt nee= ds we had to change many things. As of a year ago some tests had to be= modified, otherwise we might start tuning for the wrong goals. The tit= le of this email and the article is NetBSD beats FreeBSD in performance sho= uld be changed to NetBSD beats FreeBSD on comprehensive microbenchmark suit= e. I don't remember who it was, but someone posted results of real serv= er tests. These were in favor of FreeBSD. I doubt that much changed sin= ce last year but who knows ... Maybe NetBSD is better, but this is not som= ething that could be concluded from a microbenchmark. IMHO. igor.&#= 10; -----Original Message----- From: Hubert Feyrer To: freebsd-performance@FreeBSD.org, freebsd-advocaccy@Fre= eBSD.org Date: Thu, 6 Jan 2005 02:19:37 +0100 (CET) Subject: Benchm= ark: NetBSD 2.0 be > ats FreeBSD 5.3 in server performance > > > Gregory M= cGarry asked me to host and advertize this paper for him: > > Abst= ract: ``With the recent releases of NetBSD 2.0 and FreeBSD 5.3 operating &#= 10;> system, many new and exciting features have been implemented. Both cri= ticism > and commendation on performance, reliability and scalability = have been directed > towards these releases. > > This paper p= resents a suite of benchmarks and results for comparing the > performa= nce of these operating systems. The benchmarks target core operating >= system functionality, server scalability and thread implementation. These = > benchmarks are useful server-based criteria for demanding applicatio= ns such as > loaded webservers, databases, and voice-over-IP (VoIP) me= dia relays. The > results indicate that NetBSD has surpassed FreeBSD i= n performance on nearly > every benchmark and is poised to grab the ti= tle of the best o > perating system > for the server environment.'' > > URL: h= ttp://www.feyrer.de/NetBSD/gmcgarry/ > > > - Hubert > &= #10;> -- > NetBSD - Free AND Open! (And of course secure, portabl= e, yadda yadda) > ________________________________________________ = > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.o= rg/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail= to "freebsd-performance-unsubscribe@freebsd.org" > http= ://Mail.Ru - ?????? ????? ? ?????????????? ??????? ????????? ?????! > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd= .org" --=20 Bosko Milekic bmilekic@technokratis.com bmilekic@FreeBSD.org From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 11:29:47 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9995C16A4CE for ; Thu, 6 Jan 2005 11:29:47 +0000 (GMT) Received: from mongers.org (miracle.mongers.org [193.162.142.71]) by mx1.FreeBSD.org (Postfix) with SMTP id D1FEC43D1D for ; Thu, 6 Jan 2005 11:29:46 +0000 (GMT) (envelope-from jlouis@mongers.org) Received: (qmail 21097 invoked by uid 1030); 6 Jan 2005 11:29:44 -0000 From: "Jesper Louis Andersen" Date: Thu, 6 Jan 2005 12:29:44 +0100 To: Igor Shmukler Message-ID: <20050106112944.GA30825@miracle.mongers.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2i cc: Hubert Feyrer cc: freebsd-performance@FreeBSD.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 11:29:47 -0000 Quoting Igor Shmukler (shmukler@mail.ru): [Microbenchmarks doesn't work] Do you have a reference to the real-world benchmark you mention? It could be interesting to redo the work with a NetBSD 2.0 kernel, now it is official. -- jlouis From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 11:41:56 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 445B816A4CE for ; Thu, 6 Jan 2005 11:41:56 +0000 (GMT) Received: from mongers.org (miracle.mongers.org [193.162.142.71]) by mx1.FreeBSD.org (Postfix) with SMTP id 64B1343D46 for ; Thu, 6 Jan 2005 11:41:55 +0000 (GMT) (envelope-from jlouis@mongers.org) Received: (qmail 2400 invoked by uid 1030); 6 Jan 2005 11:41:54 -0000 From: "Jesper Louis Andersen" Date: Thu, 6 Jan 2005 12:41:54 +0100 To: Bosko Milekic Message-ID: <20050106114154.GB30825@miracle.mongers.org> References: <20050106031847.GA22306@technokratis.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20050106031847.GA22306@technokratis.com> User-Agent: Mutt/1.4.2i cc: freebsd-performance@freebsd.org cc: Hubert Feyrer cc: Igor Shmukler Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 11:41:56 -0000 Quoting Bosko Milekic (bmilekic@technokratis.com): > Not to mention, FreeBSD5 has yet to be micro-optimized. How about some > scalability benchmarks on multi-CPU machines? The original post > (particular since it was sent to -advocacy) is FUD. How many CPU's do you have in mind? I would not expect FreeBSD to outperform NetBSD by much for a 2-CPU box with a typical server workload with typical programs that does not even know to take advantage of a ''superior'' threading model. For a computer with 4-8 CPU's the advantage might be much bigger, but I have not yet seen any benchmarks targetting computers with that number of CPUs. Partially because people does not yet have access to such computers, partially because most people doesn't care about that kind of scalability. But this is speculation. I would like to see perfarmonce benchmarks for your scenario as well. I disagree that the original post is entirely FUD. While the conclusion is subjective, fact is that at the particular mix of microbenchmarks shows NetBSD faster than FreeBSD. I am wondering if that is the price you pay on single-cpu boxes to gain speed at the SMP boxes. And if this is true the question becomes if fine-grained locking is worth the implementation time when most computers are still single-cpu (Yes, I know this can change rapidly with the newer CPU types). -- jlouis From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 12:17:00 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id ABCE516A4CE for ; Thu, 6 Jan 2005 12:17:00 +0000 (GMT) Received: from multiplay.co.uk (www1.multiplay.co.uk [212.42.16.7]) by mx1.FreeBSD.org (Postfix) with ESMTP id EC9BA43D46 for ; Thu, 6 Jan 2005 12:16:59 +0000 (GMT) (envelope-from killing@multiplay.co.uk) Received: from stevenp4 ([193.123.241.40]) by multiplay.co.uk (multiplay.co.uk [212.42.16.7]) (MDaemon.PRO.v7.2.2.R) with ESMTP id md50000851329.msg for ; Thu, 06 Jan 2005 12:07:29 +0000 Message-ID: <03d501c4f3e9$92daf850$7f06000a@int.mediasurface.com> From: "Steven Hartland" To: "Jesper Louis Andersen" , "Igor Shmukler" References: <20050106112944.GA30825@miracle.mongers.org> Date: Thu, 6 Jan 2005 12:16:38 -0000 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=original Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.2180 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180 X-Spam-Processed: multiplay.co.uk, Thu, 06 Jan 2005 12:07:29 +0000 (not processed: message from valid local sender) X-MDRemoteIP: 193.123.241.40 X-Return-Path: killing@multiplay.co.uk X-MDaemon-Deliver-To: freebsd-performance@FreeBSD.org X-MDAV-Processed: multiplay.co.uk, Thu, 06 Jan 2005 12:07:32 +0000 cc: Hubert Feyrer cc: freebsd-performance@FreeBSD.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 12:17:00 -0000 I'd be interested to the know the tools / benchmark apps used here. anyone got any links? Steve / K ----- Original Message ----- From: "Jesper Louis Andersen" > Quoting Igor Shmukler (shmukler@mail.ru): > > [Microbenchmarks doesn't work] > > Do you have a reference to the real-world benchmark you mention? It > could be interesting to redo the work with a NetBSD 2.0 kernel, now > it is official. ================================================ This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone (023) 8024 3137 or return the E.mail to postmaster@multiplay.co.uk. From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 13:23:35 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9A32316A4CE for ; Thu, 6 Jan 2005 13:23:35 +0000 (GMT) Received: from smtp100.rog.mail.re2.yahoo.com (smtp100.rog.mail.re2.yahoo.com [206.190.36.78]) by mx1.FreeBSD.org (Postfix) with SMTP id 081BD43D2D for ; Thu, 6 Jan 2005 13:23:35 +0000 (GMT) (envelope-from mikej@rogers.com) Received: from unknown (HELO cpe000103d44c07-cm000f9f7ae88c.cpe.net.cable.rogers.com) (mikej@69.193.222.195 with login) by smtp100.rog.mail.re2.yahoo.com with SMTP; 6 Jan 2005 13:23:34 -0000 Received: from 207.219.213.163 (proxying for unknown) (SquirrelMail authenticated user mikej); by cpe000103d44c07-cm000f9f7ae88c.cpe.net.cable.rogers.com with HTTP; Thu, 6 Jan 2005 08:23:27 -0500 (EST) Message-ID: <49877.207.219.213.163.1105017807.squirrel@207.219.213.163> Date: Thu, 6 Jan 2005 08:23:27 -0500 (EST) From: "Mike Jakubik" To: freebsd-performance@FreeBSD.org User-Agent: SquirrelMail/1.4.3a X-Mailer: SquirrelMail/1.4.3a MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 13:23:35 -0000 Jesper Louis Andersen said: > I disagree that the original post is entirely FUD. While the conclusion is subjective, fact is that at the particular mix of microbenchmarks shows NetBSD faster than FreeBSD. I am wondering if that is the price you pay on single-cpu boxes to gain speed at the SMP boxes. And if this is true the question becomes if fine-grained locking is worth the implementation time when most computers are still single-cpu (Yes, I know this can change rapidly with the newer CPU types). It will change. Both AMD and Intel will be releasing dual core CPUs this year. This seems to be the future for most CPUs. From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 14:01:06 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 73C0C16A4CE for ; Thu, 6 Jan 2005 14:01:06 +0000 (GMT) Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8C53E43D54 for ; Thu, 6 Jan 2005 14:01:05 +0000 (GMT) (envelope-from robert@fledge.watson.org) Received: from fledge.watson.org (localhost [127.0.0.1]) by fledge.watson.org (8.13.1/8.13.1) with ESMTP id j06DvEZQ096450; Thu, 6 Jan 2005 08:57:14 -0500 (EST) (envelope-from robert@fledge.watson.org) Received: from localhost (robert@localhost)j06DvEUj096447; Thu, 6 Jan 2005 13:57:14 GMT (envelope-from robert@fledge.watson.org) Date: Thu, 6 Jan 2005 13:57:14 +0000 (GMT) From: Robert Watson X-Sender: robert@fledge.watson.org To: Jesper Louis Andersen In-Reply-To: <20050106114154.GB30825@miracle.mongers.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-performance@freebsd.org cc: Bosko Milekic cc: Igor Shmukler cc: Hubert Feyrer Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 14:01:06 -0000 On Thu, 6 Jan 2005, Jesper Louis Andersen wrote: > But this is speculation. I would like to see perfarmonce benchmarks for > your scenario as well. Since the performance optimization on FreeBSD for the last few years, with features like SMPng, libpthread, and UMA, has been focussed on macro performance not micro performance, it's not surprising the micro performance requires tuning. However, there is lots of on-going work on this front so I'd expect to see continued improvement in the immediate (5.4) life time. There are a number of optimizations in 6.x that are on the merge path for 5.x that will directly impact the results in these measurements -- in particular, what is clearly a bug in the way mutexes are released on UP kernels that adds almost a hundred cycles to every mutex release operation. This was identified in my micro-benchmarking shortly after the release of 5.3, and may play a substantial part in the posted results, especially for the very micro benchmarks that involve kernel memory allocation. Hopefully for 5.4 we'll also have a move to soft critical sections protecting UMA memory caches, which should eliminate mutex operations from common case memory allocation making a further benefit. Those changes are not yet in CVS HEAD, but are in Perforce and show nice benefits in micro-benchmarks involving kernel object allocation, such as socket allocation, etc. > I disagree that the original post is entirely FUD. While the conclusion > is subjective, fact is that at the particular mix of microbenchmarks > shows NetBSD faster than FreeBSD. I am wondering if that is the price > you pay on single-cpu boxes to gain speed at the SMP boxes. And if this > is true the question becomes if fine-grained locking is worth the > implementation time when most computers are still single-cpu (Yes, I > know this can change rapidly with the newer CPU types). I think the post Bosko is referring to is indeed almost entirely FUD, since it consists of one line of useful content ("Look, a set of decent-looking set of microbenchmark results") and the rest is ("I hate these specific FreeBSD developers, why do we let committers design the system!"). I.e., the posted report was simply an excuse for someone to chew out the FreeBSD developers. That doesn't invalidate the report, but it does suggest that the interpretation presented in that post might be a little unbalanced. Obviously, we need to read and understand the report, and act on areas where we can improve. Regarding SMP or not -- the path the FreeBSD Project has taken (and this choice was before I was really all that involved, to be honest) was a re-architecture of the kernel to improve performance, scalability, and structure via a movement to a parallelizable, preemptible, threaded kernel. I think this is the right architecture to move to, as it not only improves performance and scalability, but it also closes a lot of existing race conditions in the kernel that only became more exposed as threading and SMP became more predominant. This has had a lot of performance benefits, but comes with initial costs that aren't all immediately offset by initial benefits. Now that this model is largely adopted, we'll see a nice increase in benefits over time -- i.e., it was an investment. Obviously, people can and will disagree about the nature of the investment, the time for payoff, etc, but I think it's easy to argue there have been some very strong benefits so far. Likewise, it's possible to argue that the exact path by which it was implemented could have been better -- i.e., if the dotcom crash hadn't happened at the wrong moment leaving us with far fewer developers working on it than hoped. However, we've accomplished a lot, especially given the available resources. Immediate and measurable performance from the new architecture is now the primary thrust of the netperf work, having gotten the initial cut working, so we should see some large gains in that area, having an immediate effect on both micro-benchmarks and macro-benchmarks. More on SMP generally -- as other posts argue, I think you'll see SMP become more and more a part of "out of the box" systems over the next couple of years, especially for server class hardware. It's quite hard to buy decent server equipment from Intel that doesn't have at least HTT today. It is very important that we re-optimize UP having adopted the SMPng approach, and there's a substantial amount of work going into that, but SMP is an increasing reality that the FreeBSD Project has been addressing head-on for several years. It required a lot of work over that time, but as a result of that investment, we will be ready for the next generation of systems where SMP is no longer an option. So I'd look to immediate performance improvements in 5.x-STABLE for UP and SMP over the next few months leading up to 5.4. We should all bug Stephan Uphoff and John Baldwin to get the critical section stuff in the tree so I can merge in my UMA changes, and make sure that the UP mutex optimizations are also in RELENG_5 :-). Robert N M Watson From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 05:22:47 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 769FC16A4CE for ; Thu, 6 Jan 2005 05:22:47 +0000 (GMT) Received: from web53902.mail.yahoo.com (web53902.mail.yahoo.com [206.190.36.212]) by mx1.FreeBSD.org (Postfix) with SMTP id 049C043D3F for ; Thu, 6 Jan 2005 05:22:47 +0000 (GMT) (envelope-from stheg_olloydson@yahoo.com) Received: (qmail 47982 invoked by uid 60001); 6 Jan 2005 05:22:46 -0000 Comment: DomainKeys? See http://antispam.yahoo.com/domainkeys DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; b=Qj4fnXdtzh3Fyw7kYD4A9/rc+pyhlNlWzji2ayJkW/1UecJBnheW31Wp3AzBISMnvsNhHzbmzjhCfej5g38SLlo6EsVfFTxZzXgtrrfXn3RFyW0qyxdCsXdYFISo8Aa7iZvrfNSkocOpc6D2Vn7NN7xwpLyo30Fxi3pXhGiWWsA= ; Message-ID: <20050106052246.47980.qmail@web53902.mail.yahoo.com> Received: from [68.18.48.71] by web53902.mail.yahoo.com via HTTP; Wed, 05 Jan 2005 21:22:46 PST Date: Wed, 5 Jan 2005 21:22:46 -0800 (PST) From: stheg olloydson To: hubert@feyrer.de, g.mcgarry@ieee.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailman-Approved-At: Thu, 06 Jan 2005 14:31:35 +0000 cc: freebsd-performance@FreeBSD.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 05:22:47 -0000 it was said: Gregory McGarry asked me to host and advertize this paper for him: Abstract: ``With the recent releases of NetBSD 2.0 and FreeBSD 5.3 operating system, many new and exciting features have been implemented. Both criticism and commendation on performance, reliability and scalability have been directed towards these releases. This paper presents a suite of benchmarks and results for comparing the performance of these operating systems. The benchmarks target core operating system functionality, server scalability and thread implementation. URL: http://www.feyrer.de/NetBSD/gmcgarry/ --------------- Hello, I read the "paper" with which you trolled a couple of FreeBSD lists. Here, quoting the "paper" is what stands out to me: Abstract: "These benchmarks are useful server-based criteria for demanding applications such as loaded webservers, databases, and voice-over-IP (VoIP) media relays. The results indicate that NetBSD has surpassed FreeBSD in performance on nearly every benchmark and is poised to grab the title of the best operating system for the server environment." System-call overhead "The results in Table 1 shows that NetBSD 2.0 marginally out-performs FreeBSD 5.3." Context-switch time "The results in Table 1 shows that NetBSD 2.0 marginally outperforms FreeBSD 5.3." Process lifecycle "The results in Table 1 indicate that NetBSD 2.0 marginally outperforms FreeBSD 5.3." Forking new processes "The FreeBSD kernel has an access time which scales linearly with the number of system processes. There are also many occasions when the access-time is very fast, resembling a constant access time." Memory-mapped files "The total of both benchmarks indicate that for a single mapping and subsequent access, FreeBSD shows a 38% performance improvement over NetBSD." Socket creation scalability "Neither NetBSD nor FreeBSD shows scalability problems." Binding addresses to sockets "This result indicates the FreeBSD scales better for binding addresses to sockets." Thread creation benchmark "For less than 250 threads, the time to create a thread is better in NetBSD than FreeBSD. For more than 250 threads, the thread creation time increases as the number of threads increases. Of particular concern, the relationship is not linear for the number of threads. Although one thousand threads is ample for most multi-threaded applications, the poor scalability may be a problem for some applications." Thread lifecycle, condition variables and mutexes "The results of these benchmarks for the basic POSIX thread primitives clearly shows that the NetBSD thread implementation contains many impressive optimizations." Thread context-switch time "The creation of the first thread incurs a significant penalty on NetBSD. However, the time to complete the game is significantly higher for FreeBSD over NetBSD. This is due to increased latency in the thread lifecycle and the much longer mutex acquire time for FreeBSD over NetBSD." Conclusions "Significant performance improvements are obviously visible in the thread implementation....Microbenchmarks are not always the best indicators to make judgments on the overall performance of one operating system over another...For many applications, the results presented in the paper may never affect performance. For others, the scalability of the operating system may simply not permit the application to run suitably." Clearly, the claims in the Abstract are not supported by the tests. The author even points out the inapplicability of the benchmarks that he ran. But even if they were, the author points out that only in the case of threading performance does NetBSD do appreciably better than FreeBSD. In all other tests, FreeBSD either does better, as well as, or only "marginally" worse than NetBSD. So what this should have been called is, "Meaningless Tests Show NetBSD 2.0 is Finally About as Good as FreeBSD 5.3." I want the 20 minutes that I spent reading this "paper" and responding to it back. If you are unwilling or unable to return the piece of my life you caused me to waste, I demand to personally apologise to everyone who reads not only the post, but the "paper", as well. stheg __________________________________ Do you Yahoo!? Yahoo! Mail - Easier than ever with enhanced search. Learn more. http://info.mail.yahoo.com/mail_250 From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 10:42:39 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id B170616A4CE; Thu, 6 Jan 2005 10:42:39 +0000 (GMT) Received: from mail.esiee.fr (mail.esiee.fr [147.215.1.3]) by mx1.FreeBSD.org (Postfix) with ESMTP id ECCC943D2D; Thu, 6 Jan 2005 10:42:38 +0000 (GMT) (envelope-from f.bonnet@esiee.fr) Received: from localhost.esiee.fr (localhost.esiee.fr [127.0.0.1]) by mail.esiee.fr (Postfix) with ESMTP id 590C93658E7; Thu, 6 Jan 2005 11:42:35 +0100 (CET) Received: from mail.esiee.fr (localhost.esiee.fr [127.0.0.1]) by localhost.esiee.fr (VaMailArmor-2.0.1.16) id 83493-130B8B94; Thu, 06 Jan 2005 11:42:35 +0100 Received: from [147.215.1.13] (desolation.esiee.fr [147.215.1.13]) by mail.esiee.fr (Postfix) with ESMTP id 064213658BE; Thu, 6 Jan 2005 11:42:32 +0100 (CET) Message-ID: <41DD1615.7050908@esiee.fr> Date: Thu, 06 Jan 2005 11:42:29 +0100 From: Frank Bonnet User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hubert Feyrer References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-AntiVirus: checked by Vexira MailArmor (version: 2.0.1.16; VAE: 6.29.0.5; VDF: 6.29.0.51; host: mail.esiee.fr) X-Mailman-Approved-At: Thu, 06 Jan 2005 14:31:35 +0000 cc: freebsd-advocacy@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 10:42:39 -0000 Hubert Feyrer wrote: > > Gregory McGarry asked me to host and advertize this paper for him: > > Abstract: ``With the recent releases of NetBSD 2.0 and FreeBSD 5.3 > operating system, many new and exciting features have been implemented. > Both criticism and commendation on performance, reliability and > scalability have been directed towards these releases. Does NetBSD is able to use an external LDAP server to authenticate users that is use nss_ldap and pam_ldap ? Thanks a lot -- Cordialement/Regards Frank Bonnet From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 12:19:51 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0984616A4CE; Thu, 6 Jan 2005 12:19:51 +0000 (GMT) Received: from daemon.li (daemon.li [213.203.244.86]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9C58F43D1F; Thu, 6 Jan 2005 12:19:50 +0000 (GMT) (envelope-from josef@daemon.li) Received: from localhost (localhost [127.0.0.1]) (uid 1000) by daemon.li with local; Thu, 06 Jan 2005 12:19:49 +0000 Date: Thu, 6 Jan 2005 12:19:49 +0000 From: Josef El-Rayes To: Hubert Feyrer Message-ID: <20050106121948.GA7848@daemon.li> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: User-Agent: Mutt/1.3.28i X-Mailman-Approved-At: Thu, 06 Jan 2005 14:31:35 +0000 cc: freebsd-advocacy@FreeBSD.org cc: freebsd-performance@FreeBSD.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 12:19:51 -0000 Hubert Feyrer : [...] > The results indicate that NetBSD=20 > has surpassed FreeBSD in performance on nearly every benchmark and=20 > is poised to grab the title of the best operating system for the=20 > server environment.'' I think this is a conclusion drawn too early when there has not been any comparison of each SMP implementation. No one runs a toaster as a server environment. greets, josef --=20 Josef El-Rayes (__) Email: josef@daemon.li \\\'',)=20 Web: http://daemon.li/ \/ \ ^ FreeBSD Security Team .\._/_) From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 15:09:04 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 85E0716A4CE for ; Thu, 6 Jan 2005 15:09:04 +0000 (GMT) Received: from f19.mail.ru (f19.mail.ru [194.67.57.49]) by mx1.FreeBSD.org (Postfix) with ESMTP id 39EBF43D45 for ; Thu, 6 Jan 2005 15:09:04 +0000 (GMT) (envelope-from shmukler@mail.ru) Received: from mail by f19.mail.ru with local id 1CmZG2-000BpE-00; Thu, 06 Jan 2005 18:09:02 +0300 Received: from [24.185.78.105] by mac.mail.ru with HTTP; Thu, 06 Jan 2005 18:09:02 +0300 From: Igor Shmukler To: Martin P.Hellwig Mime-Version: 1.0 X-Mailer: mPOP Web-Mail 2.19 X-Originating-IP: [24.185.78.105] Date: Thu, 06 Jan 2005 18:09:02 +0300 In-Reply-To: <41DD32C0.40608@xs4all.nl> Content-Type: text/plain; charset=koi8-r Content-Transfer-Encoding: 8bit Message-Id: cc: freebsd-performance@FreeBSD.org Subject: Re[2]: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Igor Shmukler List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 15:09:04 -0000 Hi, I do not agree with most authors regarding UP vs. MP performance. There is a benchmark suite (not first microbenchmark suite) that shows resource allocation under NetBSD is faster. That's all there is to it. Not more or less... By performance we mean an array of properties for me it starts with throughput. To measure performace we would have run things like TPC-C and other workload simulation tests. Back in a day Linux was being tuned to show better LMBench numbers. Everyone knows what happened with that. Assuming that root of a problem is number of CPUs is not warranted, unless someone actually did comparison and concluded that NetBSD resource allocation does not scale across multiple processors. Let's take things for what they are.A microbenchmark shows that some NetBSD allocators scale better. A microbenchmark cannot give any results that could be used to realistically assess an overall performance. igor. From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 16:56:48 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 19E1016A4CE for ; Thu, 6 Jan 2005 16:56:48 +0000 (GMT) Received: from rproxy.gmail.com (rproxy.gmail.com [64.233.170.203]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9111843D5F for ; Thu, 6 Jan 2005 16:56:47 +0000 (GMT) (envelope-from phil.brennan@gmail.com) Received: by rproxy.gmail.com with SMTP id 40so572611rnz for ; Thu, 06 Jan 2005 08:56:47 -0800 (PST) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references; b=b9fGAOEV/7a+AFa8vsD3LhF+w4m0trXyV5U5WCVlLbuTgWAXvsDa8SqPf9hcT66qMzBro4E1qpRGXR/VMR3Ry183Xxl4fx+uFlC0e987WZ/CH+gMtXKlAnCTe2alj4N0oN+q5V2XyKwMMo6mNKAOE8AB0I7vTl4inhWWTR3Oz34= Received: by 10.38.8.13 with SMTP id 13mr220403rnh; Thu, 06 Jan 2005 08:56:46 -0800 (PST) Received: by 10.38.179.27 with HTTP; Thu, 6 Jan 2005 08:56:46 -0800 (PST) Message-ID: Date: Thu, 6 Jan 2005 16:56:46 +0000 From: Phil Brennan In-Reply-To: <20050106121948.GA7848@daemon.li> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit References: <20050106121948.GA7848@daemon.li> cc: freebsd-advocacy@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Phil Brennan List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 16:56:48 -0000 What about the context switch time? Are there any plans to improve this, and also to reduce the number of context switches needed? On Thu, 6 Jan 2005 12:19:49 +0000, Josef El-Rayes wrote: > Hubert Feyrer : > > [...] > > > The results indicate that NetBSD > > has surpassed FreeBSD in performance on nearly every benchmark and > > is poised to grab the title of the best operating system for the > > server environment.'' > > I think this is a conclusion drawn too early when there has not been > any comparison of each SMP implementation. > No one runs a toaster as a server environment. > > greets, josef > -- > Josef El-Rayes (__) > Email: josef@daemon.li \\\'',) > Web: http://daemon.li/ \/ \ ^ > FreeBSD Security Team .\._/_) > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 17:45:38 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9D10B16A4D8 for ; Thu, 6 Jan 2005 17:45:12 +0000 (GMT) Received: from stephanie.unixdaemons.com (stephanie.unixdaemons.com [67.18.111.194]) by mx1.FreeBSD.org (Postfix) with ESMTP id 12E8F43D1F for ; Thu, 6 Jan 2005 17:45:10 +0000 (GMT) (envelope-from bmilekic@technokratis.com) Received: from stephanie.unixdaemons.com (bmilekic@localhost.unixdaemons.com [127.0.0.1])j06Hj4pw047372; Thu, 6 Jan 2005 12:45:04 -0500 (EST) Received: (from bmilekic@localhost) by stephanie.unixdaemons.com (8.13.2/8.12.1/Submit) id j06Hj3sC047371; Thu, 6 Jan 2005 12:45:03 -0500 (EST) (envelope-from bmilekic@technokratis.com) X-Authentication-Warning: stephanie.unixdaemons.com: bmilekic set sender to bmilekic@technokratis.com using -f Date: Thu, 6 Jan 2005 12:45:03 -0500 From: Bosko Milekic To: Jesper Louis Andersen Message-ID: <20050106174503.GA45214@technokratis.com> References: <20050106031847.GA22306@technokratis.com> <20050106114154.GB30825@miracle.mongers.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20050106114154.GB30825@miracle.mongers.org> User-Agent: Mutt/1.4.2.1i cc: freebsd-performance@freebsd.org cc: Hubert Feyrer cc: Igor Shmukler Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 17:45:39 -0000 On Thu, Jan 06, 2005 at 12:41:54PM +0100, Jesper Louis Andersen wrote: > Quoting Bosko Milekic (bmilekic@technokratis.com): > > > Not to mention, FreeBSD5 has yet to be micro-optimized. How about some > > scalability benchmarks on multi-CPU machines? The original post > > (particular since it was sent to -advocacy) is FUD. > > How many CPU's do you have in mind? I would not expect FreeBSD to > outperform NetBSD by much for a 2-CPU box with a typical server > workload with typical programs that does not even know to take advantage > of a ''superior'' threading model. For a computer with 4-8 CPU's the > advantage might be much bigger, but I have not yet seen any benchmarks > targetting computers with that number of CPUs. Partially because > people does not yet have access to such computers, partially because > most people doesn't care about that kind of scalability. It's not true that people do not have access to such computers. Anyway, the point is that the benchmark results were pitched on our mailing lists as being an affirmative proof that NetBSD outperforms FreeBSD in various different setups which is, I maintain, complete FUD. Measuring context switch time, micro-benchmark style, and concluding that this means that FreeBSD performs worse than NetBSD, period, is wrong. As for the multi-CPU case, it *is* important and it *is* relevant. It might not be relevant if you're looking at userland-bound processes, but there *is* lots of kernel-bound processing that goes on in both end-host and forwarding setups, and it is unjust to compare and make sweeping conclusions regarding two systems, one which is designed to handle multi-processor scalability in-kernel, and one which is not. In some respect, it is like comparing Apples with Oranges. > But this is speculation. I would like to see perfarmonce benchmarks for > your scenario as well. > > I disagree that the original post is entirely FUD. While the conclusion > is subjective, fact is that at the particular mix of microbenchmarks > shows NetBSD faster than FreeBSD. I am wondering if that is the price > you pay on single-cpu boxes to gain speed at the SMP boxes. And if this > is true the question becomes if fine-grained locking is worth the > implementation time when most computers are still single-cpu (Yes, I > know this can change rapidly with the newer CPU types). It is also a question of development _stage_. There are ways to further optimize the single CPU case, but attention has not necessarily been focused on that particular case yet, as significant architectural work remains (before further focus is diverted toward micro-optimizing). > -- > jlouis -- Bosko Milekic bmilekic@technokratis.com bmilekic@FreeBSD.org From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 17:41:26 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5C57016A4CE for ; Thu, 6 Jan 2005 17:41:26 +0000 (GMT) Received: from web53908.mail.yahoo.com (web53908.mail.yahoo.com [206.190.36.218]) by mx1.FreeBSD.org (Postfix) with SMTP id D0A0243D49 for ; Thu, 6 Jan 2005 17:41:25 +0000 (GMT) (envelope-from stheg_olloydson@yahoo.com) Received: (qmail 18036 invoked by uid 60001); 6 Jan 2005 17:41:00 -0000 Comment: DomainKeys? See http://antispam.yahoo.com/domainkeys DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; b=H/TviVG7l5RjR7HbhX/EtfgqMl8wQZriIIaPvFxnp9acBhrlbO5IXIpCtcKgIqst+UMVqeCDXdyPaIjE1l6GCzgbrdusrJSqjZKKbWJ8bxD4OgOJiHP3GpVupxzJ7yASUYEHkmJHDWivYfPI9K0dCLwofyg6MZP0fYcx5R/ISS4= ; Message-ID: <20050106174100.18034.qmail@web53908.mail.yahoo.com> Received: from [68.18.48.71] by web53908.mail.yahoo.com via HTTP; Thu, 06 Jan 2005 09:40:59 PST Date: Thu, 6 Jan 2005 09:40:59 -0800 (PST) From: stheg olloydson To: phil.brennan@gmail.com MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailman-Approved-At: Thu, 06 Jan 2005 19:30:14 +0000 cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 [allegedly] beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 17:41:26 -0000 it was said by Phil Brennan: >What about the context switch time? Are there any plans to improve >this, and also to reduce the number of context switches needed? Hello, See Robert Watson's reply to this thread. An unfortunate number of problems exist in threading and scheduling. Most are well-understood and are being worked on and 5.4 should see measurable improvement. Personally, I am more concerned with network and scheduler perfomance. I know the former is being addressed, but I don't hear anything about how work on SCHED_ULE is progressing. Regards, stheg __________________________________ Do you Yahoo!? Yahoo! Mail - 250MB free storage. Do more. Manage less. http://info.mail.yahoo.com/mail_250 From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 18:11:43 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 07BD816A4CF for ; Thu, 6 Jan 2005 18:11:43 +0000 (GMT) Received: from rproxy.gmail.com (rproxy.gmail.com [64.233.170.194]) by mx1.FreeBSD.org (Postfix) with ESMTP id 779C343D48 for ; Thu, 6 Jan 2005 18:11:42 +0000 (GMT) (envelope-from sander.vesik@gmail.com) Received: by rproxy.gmail.com with SMTP id y7so28562rne for ; Thu, 06 Jan 2005 10:11:41 -0800 (PST) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references; b=aeMfY6OgN6ZTqp8K1v94nE7ZubX099IoI0BYh5Nc02zikxQby0y8qohft3auORoSuMOBDk1XqX/aGVEvBpJyX/2Ylh5GdE897VWvG/OpEr+x8IFSc2QDSQtycYhlGR+dibevrTntzF6f7sAxkEQO1KKJRLJuwlO+X+H654mQX0A= Received: by 10.38.208.29 with SMTP id f29mr381130rng; Thu, 06 Jan 2005 10:11:41 -0800 (PST) Received: by 10.38.66.66 with HTTP; Thu, 6 Jan 2005 10:11:41 -0800 (PST) Message-ID: Date: Thu, 6 Jan 2005 20:11:41 +0200 From: Sander Vesik To: Josef El-Rayes In-Reply-To: <20050106121948.GA7848@daemon.li> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit References: <20050106121948.GA7848@daemon.li> X-Mailman-Approved-At: Thu, 06 Jan 2005 19:30:14 +0000 cc: freebsd-advocacy@freebsd.org cc: Hubert Feyrer cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Sander Vesik List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 18:11:43 -0000 On Thu, 6 Jan 2005 12:19:49 +0000, Josef El-Rayes wrote: > Hubert Feyrer : > > [...] > > > The results indicate that NetBSD > > has surpassed FreeBSD in performance on nearly every benchmark and > > is poised to grab the title of the best operating system for the > > server environment.'' > > I think this is a conclusion drawn too early when there has not been > any comparison of each SMP implementation. > No one runs a toaster as a server environment. No but many people run servers on single CPU machines and performance on those matters too. Just because a benchmark result is not what you might like it to tell you doesn't mean its not valid or that it doesn't highlight valid concerns. For example on the process creation benchmark (and yes, it is a valid and interesting benchmark, even in uniprocessor case), its clear that both systems exhibit a split behaviour where in some processes are creating in some linear minimal time and others scale lineraily with number of processes. It just happens that in case of freebsd the majority appear to follow the linear case. There are also cases where FreeBSD is clearly ahead, which is good. > > greets, josef > -- > Josef El-Rayes (__) > Email: josef@daemon.li \\\'',) > Web: http://daemon.li/ \/ \ ^ > FreeBSD Security Team .\._/_) From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 19:40:35 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C5AC516A4D2 for ; Thu, 6 Jan 2005 19:40:35 +0000 (GMT) Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4803943D1F for ; Thu, 6 Jan 2005 19:40:35 +0000 (GMT) (envelope-from robert@fledge.watson.org) Received: from fledge.watson.org (localhost [127.0.0.1]) by fledge.watson.org (8.13.1/8.13.1) with ESMTP id j06JanFb002832; Thu, 6 Jan 2005 14:36:49 -0500 (EST) (envelope-from robert@fledge.watson.org) Received: from localhost (robert@localhost)j06Janm3002829; Thu, 6 Jan 2005 19:36:49 GMT (envelope-from robert@fledge.watson.org) Date: Thu, 6 Jan 2005 19:36:49 +0000 (GMT) From: Robert Watson X-Sender: robert@fledge.watson.org To: stheg olloydson In-Reply-To: <20050106174100.18034.qmail@web53908.mail.yahoo.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: phil.brennan@gmail.com cc: freebsd-performance@freebsd.org Subject: Re: Benchmark: NetBSD 2.0 [allegedly] beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 19:40:35 -0000 On Thu, 6 Jan 2005, stheg olloydson wrote: > it was said by Phil Brennan: > > >What about the context switch time? Are there any plans to improve > >this, and also to reduce the number of context switches needed? > > See Robert Watson's reply to this thread. An unfortunate number of > problems exist in threading and scheduling. Most are well-understood and > are being worked on and 5.4 should see measurable improvement. > Personally, I am more concerned with network and scheduler perfomance. > I know the former is being addressed, but I don't hear anything about > how work on SCHED_ULE is progressing. FWIW, one of the reasons that there hasn't been as much interest in SCHED_ULE lately is likely that several of the features previously only present in SCHED_ULE are now also present in SCHED_4BSD -- for example, making more effective uses of IPIs in reducing latency during inter-process communication across processors. While SCHED_ULE does contain a number of interesting things not present in SCHED_4BSD, the 4BSD scheduler has hardly gone un-improved in that time. However, Jeff Robserson does seem to have picked up recently on both VFS SMP locking and ULE. The scheduler tracing and visualization tools he committed a couple of weeks ago are really quite neat tools. Robert N M Watson From owner-freebsd-performance@FreeBSD.ORG Thu Jan 6 19:49:56 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5450816A4CE for ; Thu, 6 Jan 2005 19:49:56 +0000 (GMT) Received: from mail.gmx.net (mail.gmx.net [213.165.64.20]) by mx1.FreeBSD.org (Postfix) with SMTP id 4136443D1F for ; Thu, 6 Jan 2005 19:49:55 +0000 (GMT) (envelope-from michaelnottebrock@gmx.net) Received: (qmail invoked by alias); 06 Jan 2005 19:49:53 -0000 Received: from p3EE26E92.dip.t-dialin.net (EHLO lofi.dyndns.org) (62.226.110.146) by mail.gmx.net (mp023) with SMTP; 06 Jan 2005 20:49:53 +0100 X-Authenticated: #443188 Received: from kiste.my.domain (kiste.my.domain [192.168.8.4]) (authenticated bits=0) by lofi.dyndns.org (8.12.10/8.12.10) with ESMTP id j06Jnk4o033346 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NO); Thu, 6 Jan 2005 20:49:47 +0100 (CET) (envelope-from michaelnottebrock@gmx.net) From: Michael Nottebrock To: freebsd-performance@freebsd.org Date: Thu, 6 Jan 2005 20:49:36 +0100 User-Agent: KMail/1.7.2 References: <20050106174100.18034.qmail@web53908.mail.yahoo.com> In-Reply-To: <20050106174100.18034.qmail@web53908.mail.yahoo.com> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart2623881.i67rEbNr1u"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit Message-Id: <200501062049.40071.michaelnottebrock@gmx.net> X-Virus-Scanned: by amavisd-new X-Y-GMX-Trusted: 0 cc: phil.brennan@gmail.com cc: stheg olloydson Subject: Re: Benchmark: NetBSD 2.0 [allegedly] beats FreeBSD 5.3 in server performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Jan 2005 19:49:56 -0000 --nextPart2623881.i67rEbNr1u Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline On Thursday, 6. January 2005 18:40, stheg olloydson wrote: > I know the former is being addressed, but I don't hear anything about > how work on SCHED_ULE is progressing. It's at the very-least unbroken again in -CURRENT. =2D-=20 ,_, | Michael Nottebrock | lofi@freebsd.org (/^ ^\) | FreeBSD - The Power to Serve | http://www.freebsd.org \u/ | K Desktop Environment on FreeBSD | http://freebsd.kde.org --nextPart2623881.i67rEbNr1u Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (FreeBSD) iD8DBQBB3ZZUXhc68WspdLARAty2AKCEobEcpoUjMuutGZwgP7YchFhKhACeNt4M clWb7+rO2HY6abmqTGsm9ug= =/oIP -----END PGP SIGNATURE----- --nextPart2623881.i67rEbNr1u--