From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 10:16:22 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 2696016A418 for ; Mon, 10 Sep 2007 10:16:22 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from melon.pingpong.net (melon.pingpong.net [195.178.174.161]) by mx1.freebsd.org (Postfix) with ESMTP id BCD0F13C4A5 for ; Mon, 10 Sep 2007 10:16:21 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from localhost (localhost.pingpong.net [127.0.0.1]) by melon.pingpong.net (Postfix) with ESMTP id CFE5250854 for ; Mon, 10 Sep 2007 11:48:53 +0200 (CEST) Received: from melon.pingpong.net ([127.0.0.1]) by localhost (melon.pingpong.net [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 94538-01-33 for ; Mon, 10 Sep 2007 11:48:53 +0200 (CEST) Received: from [192.168.1.187] (unknown [213.136.40.204]) by melon.pingpong.net (Postfix) with ESMTP id A3CC35084D for ; Mon, 10 Sep 2007 11:48:53 +0200 (CEST) Date: Mon, 10 Sep 2007 11:48:53 +0200 From: Palle Girgensohn To: freebsd-performance@freebsd.org Message-ID: X-Mailer: Mulberry/4.0.8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; FORMAT=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-Virus-Scanned: by amavisd-new at pingpong.net Subject: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 10:16:22 -0000 Hi, We are looking at getting a server for running postgresql. Only postgresql, a dedicated machine. Since we know FreeBSD very well, we plan on using it as the OS. We have an offer for an IBM server, x3650, with 2 * DC Intel Xeon 5160, Raid with two clusters, one for database and one for xlog. Loads and loads of RAM. Now, I hear rumors that AMD is to be preferred over Intel for performance reasons. Any comments on that? Are we better off getting AMD processors, when running Unix? Regards, Palle From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 10:46:33 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 6115E16A420 for ; Mon, 10 Sep 2007 10:46:33 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from melon.pingpong.net (melon.pingpong.net [195.178.174.161]) by mx1.freebsd.org (Postfix) with ESMTP id 016AA13C47E for ; Mon, 10 Sep 2007 10:46:32 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from localhost (localhost.pingpong.net [127.0.0.1]) by melon.pingpong.net (Postfix) with ESMTP id A1F3350855; Mon, 10 Sep 2007 12:46:31 +0200 (CEST) Received: from melon.pingpong.net ([127.0.0.1]) by localhost (melon.pingpong.net [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 02672-01-15; Mon, 10 Sep 2007 12:46:31 +0200 (CEST) Received: from [192.168.1.187] (unknown [213.136.40.204]) by melon.pingpong.net (Postfix) with ESMTP id 0D28D50854; Mon, 10 Sep 2007 12:46:31 +0200 (CEST) Date: Mon, 10 Sep 2007 12:46:30 +0200 From: Palle Girgensohn To: Erich Dollansky Message-ID: <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> In-Reply-To: <46E51D32.5070803@pacific.net.sg> References: <46E51D32.5070803@pacific.net.sg> X-Mailer: Mulberry/4.0.8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable Content-Disposition: inline X-Virus-Scanned: by amavisd-new at pingpong.net Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 10:46:33 -0000 --On m=E5ndag, september 10, 2007 18.32.18 +0800 Erich Dollansky=20 wrote: > Hi, > > Palle Girgensohn wrote: >> >> Now, I hear rumors that AMD is to be preferred over Intel for >> performance reasons. Any comments on that? Are we better off getting AMD >> processors, when running Unix? >> > AMD announced today the new quad core. Their benchmarks are promising. > > But not the dual core versions. OK, thanks! But this means they can't deliver very soon, I guess, so I'd=20 better stick to woodcrest if I want to buy something soon. Thanks for your input! Palle From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 10:51:42 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 92B7F16A41B for ; Mon, 10 Sep 2007 10:51:42 +0000 (UTC) (envelope-from kometen@gmail.com) Received: from nf-out-0910.google.com (nf-out-0910.google.com [64.233.182.186]) by mx1.freebsd.org (Postfix) with ESMTP id 08C8113C48D for ; Mon, 10 Sep 2007 10:51:41 +0000 (UTC) (envelope-from kometen@gmail.com) Received: by nf-out-0910.google.com with SMTP id k4so770749nfd for ; Mon, 10 Sep 2007 03:51:40 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=YjcAt7iTqHAcfWwDJU9EpJG+LcHH/2u63jCklrms5G4=; b=t06cHfT9hWab1uOwz7/e7lniQcSmg3KUDzLWHqj+ta80B0xcLuPMdhN4bIPrTUOLeaJ13YngPLOJ9IFEc9G8n71XTRPxP5xzbV4hiNdY82DsWE4sVTB5Bd5NZq9eA63K+Y5SGfpDsHoiYek2dIRJm4nJCnIswI7px4v7Tnf+NY0= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=hWQTEFVLNTDJlo3RttbXHtNvkRoFz0M+mJCpatloyv+qDQ8eop1FZb+5crybRrhhvq1Tx3SlRiwFzGnyn9ELgmtfpYewr5vlIkBo8jaq+R9fWUfRiTVWWJ4zuD8AaBwNMH4xyEbaQLmY1VsPO2MuVhm74kHxxFpsa9O4iwPCFgA= Received: by 10.78.118.5 with SMTP id q5mr1848334huc.1189419926734; Mon, 10 Sep 2007 03:25:26 -0700 (PDT) Received: by 10.78.146.10 with HTTP; Mon, 10 Sep 2007 03:25:26 -0700 (PDT) Message-ID: Date: Mon, 10 Sep 2007 12:25:26 +0200 From: "Claus Guttesen" To: "Palle Girgensohn" In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 10:51:42 -0000 > We are looking at getting a server for running postgresql. Only postgresql, > a dedicated machine. Since we know FreeBSD very well, we plan on using it > as the OS. > > We have an offer for an IBM server, x3650, with 2 * DC Intel Xeon 5160, > Raid with two clusters, one for database and one for xlog. Loads and loads > of RAM. > > Now, I hear rumors that AMD is to be preferred over Intel for performance > reasons. Any comments on that? Are we better off getting AMD processors, > when running Unix? Not atm. The newer xeon's (woodcrest and friends) are no longer zzeonns, but are rather snappy and outperforms opterons. You can look at http://tweakers.net/reviews/657/6 as one example. -- regards Claus When lenity and cruelty play for a kingdom, the gentlest gamester is the soonest winner. Shakespeare From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 10:54:29 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 846B616A469 for ; Mon, 10 Sep 2007 10:54:29 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: from smtpgate1.pacific.net.sg (smtpgate1.pacific.net.sg [203.120.90.31]) by mx1.freebsd.org (Postfix) with SMTP id A171A13C46C for ; Mon, 10 Sep 2007 10:54:28 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: (qmail 6935 invoked from network); 10 Sep 2007 10:54:26 -0000 Received: from bb121-7-106-120.singnet.com.sg (HELO P2120.somewherefaraway.com) (oceanare@121.7.106.120) by smtpgate1.pacific.net.sg with ESMTPA; 10 Sep 2007 10:54:26 -0000 Message-ID: <46E52260.60000@pacific.net.sg> Date: Mon, 10 Sep 2007 18:54:24 +0800 From: Erich Dollansky User-Agent: Thunderbird 2.0.0.6 (X11/20070826) MIME-Version: 1.0 To: Palle Girgensohn References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> In-Reply-To: <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 10:54:29 -0000 Hi, Palle Girgensohn wrote: > > >>> Now, I hear rumors that AMD is to be preferred over Intel for >>> performance reasons. Any comments on that? Are we better off getting AMD >>> processors, when running Unix? >>> >> AMD announced today the new quad core. Their benchmarks are promising. >> >> But not the dual core versions. > > OK, thanks! But this means they can't deliver very soon, I guess, so I'd > better stick to woodcrest if I want to buy something soon. > they are delivering already. Those German guys have had real silicon to test with. I have no idea how long it will take to get one if your name is not Sun or Cray. I think it is very simple for you. If they are not available the moment you want to buy, you will buy Intel. Erich From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 10:59:06 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 8056016A41B for ; Mon, 10 Sep 2007 10:59:06 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: from smtpgate4.pacific.net.sg (smtpgate4.pacific.net.sg [203.81.36.24]) by mx1.freebsd.org (Postfix) with SMTP id 9E1F913C46B for ; Mon, 10 Sep 2007 10:59:05 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: (qmail 20863 invoked from network); 10 Sep 2007 10:32:24 -0000 Received: from bb121-7-106-120.singnet.com.sg (HELO P2120.somewherefaraway.com) (oceanare@121.7.106.120) by smtpgate4.pacific.net.sg with ESMTPA; 10 Sep 2007 10:32:23 -0000 Message-ID: <46E51D32.5070803@pacific.net.sg> Date: Mon, 10 Sep 2007 18:32:18 +0800 From: Erich Dollansky User-Agent: Thunderbird 2.0.0.6 (X11/20070826) MIME-Version: 1.0 To: Palle Girgensohn References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 10:59:06 -0000 Hi, Palle Girgensohn wrote: > > Now, I hear rumors that AMD is to be preferred over Intel for > performance reasons. Any comments on that? Are we better off getting AMD > processors, when running Unix? > AMD announced today the new quad core. Their benchmarks are promising. But not the dual core versions. Erich From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 09:43:01 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C236916A41A for ; Mon, 10 Sep 2007 09:43:01 +0000 (UTC) (envelope-from girgen@FreeBSD.org) Received: from melon.pingpong.net (melon.pingpong.net [195.178.174.161]) by mx1.freebsd.org (Postfix) with ESMTP id 46F1213C46B for ; Mon, 10 Sep 2007 09:43:00 +0000 (UTC) (envelope-from girgen@FreeBSD.org) Received: from localhost (localhost.pingpong.net [127.0.0.1]) by melon.pingpong.net (Postfix) with ESMTP id C16D65087C; Mon, 10 Sep 2007 11:11:14 +0200 (CEST) Received: from melon.pingpong.net ([127.0.0.1]) by localhost (melon.pingpong.net [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 86122-01-21; Mon, 10 Sep 2007 11:11:14 +0200 (CEST) Received: from [192.168.1.187] (unknown [213.136.40.204]) by melon.pingpong.net (Postfix) with ESMTP id 8643750855; Mon, 10 Sep 2007 11:11:14 +0200 (CEST) Date: Mon, 10 Sep 2007 11:11:14 +0200 From: Palle Girgensohn To: freebsd-performance@freebsd.org, postgresql-performance@postgresql.org Message-ID: <475C926AAAF345AFC8B9F513@rambutan.pingpong.net> X-Mailer: Mulberry/4.0.8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-Virus-Scanned: by amavisd-new at pingpong.net X-Mailman-Approved-At: Mon, 10 Sep 2007 14:58:02 +0000 Cc: Subject: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 09:43:01 -0000 Hi, We are looking at getting a server for running postgresql. Only postgresql, a dedicated machine. Since we know FreeBSD very well, we plan on using it as the OS. We have an offer for an IBM server, x3650, with 2 * DC Intel Xeon 5160, Raid with two clusters, one for database and one for xlog. Loads and loads of RAM. Now, I hear rumors that AMD is to be preferred over Intel for performance reasons. Any comments on that? Are we better off getting AMD processors, when running Unix? Regards, Palle From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 14:29:48 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id A7EEF16A41A for ; Mon, 10 Sep 2007 14:29:48 +0000 (UTC) (envelope-from ppathiakis@eagleaccess.com) Received: from smtpout01.eagleinvsys.com (smtpout01.eagleinvsys.com [12.104.12.179]) by mx1.freebsd.org (Postfix) with ESMTP id 3E0EB13C483 for ; Mon, 10 Sep 2007 14:29:48 +0000 (UTC) (envelope-from ppathiakis@eagleaccess.com) Received: from testing.atlantisservices.com ([10.170.0.19]) by smtpout01.eagleinvsys.com with Microsoft SMTPSVC(6.0.3790.1830); Mon, 10 Sep 2007 10:18:36 -0400 Message-ID: <46E55204.20905@eagleaccess.com> Date: Mon, 10 Sep 2007 10:17:40 -0400 From: Paul Pathiakis User-Agent: Thunderbird 2.0.0.6 (X11/20070827) MIME-Version: 1.0 To: Erich Dollansky References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> In-Reply-To: <46E52260.60000@pacific.net.sg> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 10 Sep 2007 14:18:36.0741 (UTC) FILETIME=[7A27D750:01C7F3B5] X-Mailman-Approved-At: Mon, 10 Sep 2007 14:58:02 +0000 Cc: freebsd-performance@freebsd.org, Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 14:29:48 -0000 Be very, very careful in purchasing Core 2 Duo. There are major problems with the chip that have been documented across the board. Many, many people are steering clear of the chip for at least a year. This brings up a simple thought... is it better to stick with a tried and true chip at present generation or be the Microsoft oooopps... the Intel public testbed. I don't care who does what as long as my performance is good, my price is low and my TCO is even lower. Paul Erich Dollansky wrote: > Hi, > > Palle Girgensohn wrote: >> >> >>>> Now, I hear rumors that AMD is to be preferred over Intel for >>>> performance reasons. Any comments on that? Are we better off >>>> getting AMD >>>> processors, when running Unix? >>>> >>> AMD announced today the new quad core. Their benchmarks are promising. >>> >>> But not the dual core versions. >> >> OK, thanks! But this means they can't deliver very soon, I guess, so >> I'd better stick to woodcrest if I want to buy something soon. >> > they are delivering already. Those German guys have had real silicon > to test with. I have no idea how long it will take to get one if your > name is not Sun or Cray. > > I think it is very simple for you. If they are not available the > moment you want to buy, you will buy Intel. > > Erich > _______________________________________________ > 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 Mon Sep 10 18:58:49 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 38A1716A46C for ; Mon, 10 Sep 2007 18:58:49 +0000 (UTC) (envelope-from cracauer@koef.zs64.net) Received: from koef.zs64.net (koef.zs64.net [212.12.50.230]) by mx1.freebsd.org (Postfix) with ESMTP id AA15813C494 for ; Mon, 10 Sep 2007 18:58:48 +0000 (UTC) (envelope-from cracauer@koef.zs64.net) Received: from koef.zs64.net (koef.zs64.net [212.12.50.230]) by koef.zs64.net (8.14.1/8.14.1) with ESMTP id l8AIkMDS078081; Mon, 10 Sep 2007 20:46:22 +0200 (CEST) (envelope-from cracauer@koef.zs64.net) Received: (from cracauer@localhost) by koef.zs64.net (8.14.1/8.14.1/Submit) id l8AIkLKf078080; Mon, 10 Sep 2007 14:46:21 -0400 (EDT) (envelope-from cracauer) Date: Mon, 10 Sep 2007 14:46:21 -0400 From: Martin Cracauer To: Paul Pathiakis Message-ID: <20070910184621.GA77744@cons.org> References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <46E55204.20905@eagleaccess.com> User-Agent: Mutt/1.4.2.3i Cc: freebsd-performance@freebsd.org, Erich Dollansky , Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 18:58:49 -0000 For integer workloads Intel's Core2-base Xeons outperforms K8 (the old-school AMD64) by about 25-30% per clock per core. K10 seems to be 5-15% faster than K8 for integer workloads (I hope to run my benchmark suite on one thi week or weekend). However, tasks that use multiple cores and have threads on cores communicate a lot see both AMD architectures close the gap. Paul Pathiakis wrote on Mon, Sep 10, 2007 at 10:17:40AM -0400: > Be very, very careful in purchasing Core 2 Duo. There are major > problems with the chip that have been documented across the board. These have been blown out of proportion by Theo. Can you point to a demonstratable case with current Linux or BSD kernels? I also highly doubt that the AMD design is much more bug-free. Although it might. Core2 has very complicated caches, K10 is still simpler. On the other hand, if you want K8 or K10 in a modern SMP mainboard you have to live with NVidia for chipsets, and the socket F boards all have the MPC55 SATA controller, which iirc is unsupported by both BSD and Linux. MPC65 moved to AHCI so all is well - but there are no socket F boards with that SATA controller. One real advantage of Socket F over Socket 771 is that you can use normal registered DDR2, which goes dirt cheap both on ebay and new. Socket F sees mandatory FB-DIMMs, which are expensive and hot, and boards with more than 8 DIMM slots are rare. For socket F 2x8 DIMM slot boards and plenty and cheaper. Martin -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Martin Cracauer http://www.cons.org/cracauer/ FreeBSD - where you want to go, today. http://www.freebsd.org/ From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 21:09:18 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 3166416A419 for ; Mon, 10 Sep 2007 21:09:18 +0000 (UTC) (envelope-from fcash@sd73.bc.ca) Received: from smtp.sd73.bc.ca (smtp.sd73.bc.ca [142.24.13.140]) by mx1.freebsd.org (Postfix) with ESMTP id ED49913C458 for ; Mon, 10 Sep 2007 21:09:17 +0000 (UTC) (envelope-from fcash@sd73.bc.ca) Received: from localhost (localhost [127.0.0.1]) by localhost.sd73.bc.ca (Postfix) with ESMTP id ABA061A000EBB for ; Mon, 10 Sep 2007 13:44:54 -0700 (PDT) X-Virus-Scanned: Debian amavisd-new at smtp.sd73.bc.ca Received: from smtp.sd73.bc.ca ([127.0.0.1]) by localhost (smtp.sd73.bc.ca [127.0.0.1]) (amavisd-new, port 10024) with LMTP id TvhPwWdD6Zur for ; Mon, 10 Sep 2007 13:44:52 -0700 (PDT) Received: from webmail.sd73.bc.ca (webmail.sd73.bc.ca [10.10.10.17]) by smtp.sd73.bc.ca (Postfix) with ESMTP id C107A1A000EB3 for ; Mon, 10 Sep 2007 13:44:52 -0700 (PDT) Received: from 142.24.9.195 (SquirrelMail authenticated user fcash) by webmail.sd73.bc.ca with HTTP; Mon, 10 Sep 2007 13:44:53 -0700 (PDT) Message-ID: <49198.142.24.9.195.1189457093.squirrel@webmail.sd73.bc.ca> In-Reply-To: <20070910184621.GA77744@cons.org> References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> Date: Mon, 10 Sep 2007 13:44:53 -0700 (PDT) From: "Freddie Cash" To: freebsd-performance@freebsd.org User-Agent: SquirrelMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mailman-Approved-At: Mon, 10 Sep 2007 22:18:46 +0000 Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 21:09:18 -0000 On Mon, September 10, 2007 11:46 am, Martin Cracauer wrote: > On the other hand, if you want K8 or K10 in a modern SMP mainboard you > have to live with NVidia for chipsets, and the socket F boards all have > the MPC55 SATA controller, which iirc is unsupported by both BSD and > Linux. MPC65 moved to AHCI so all is well - but there are no > socket F boards with that SATA controller. Depending on what you consider "modern SMP" to mean, you can use dual-core Opterons in the Tyan Thunder K8SD-Pro (S2882) motherboard. This is a dual-socket 940 motherboard that uses the AMD-8111 chipset. All devices on the motherboard are fully supported by FreeBSD 6.2. It supports 4 GB DIMMS with 8 slots, for a total of 32 GB of RAM. For i386 setups, the BIOS can be configured to properly map around the "PCI memory hole". The onboard SATA controller is a crappy SiI, but there is an option to get it with an onboard Adaptec U320 SCSI controller (dual-channel). We use 3Ware 9550SX RAID controllers instead. There are several 32-bit/33 MHz PCI slots, a pair of 64-bit/33 MHz PCI-X slots, and a pair of 64-bit/133-MHZ PCI-X slots. The management daughter card works well as a serial console for FreeBSD. Only $450 CDN for the motherboard, and it's still available. We have over a 100 of these in service throughout the district doing file/print, xterminal (diskless), xterminal (thin-client), router, database, web, VMWare hosts, Xen hosts, and more. Hardy little things. These probably aren't the fastest motherboard on the block, and they only use DDR1 RAM, but they are fully supported under FreeBSD 6.x, Debian Linux 4.x, and Ubuntu Linux 7.04 (the only OSes we've tested). ---- Freddie Cash, LPIC-2 CCNT CCLP Helpdesk / Network Support Tech. School District 73 (250) 377-HELP [377-4357] fcash@sd73.bc.ca helpdesk@sd73.bc.ca From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 23:09:42 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 31AE516A41B for ; Mon, 10 Sep 2007 23:09:42 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 35CA013C459 for ; Mon, 10 Sep 2007 23:09:42 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: by elvis.mu.org (Postfix, from userid 1192) id 853091A4D7C; Mon, 10 Sep 2007 15:45:03 -0700 (PDT) Date: Mon, 10 Sep 2007 15:45:03 -0700 From: Alfred Perlstein To: Palle Girgensohn Message-ID: <20070910224503.GO79417@elvis.mu.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.3i Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 23:09:42 -0000 Palle, I really haven't kept pace with Intel versus AMD in a while, my understanding is that AMD is still the only 64bit game in town. For a database, the more memory you can get, the better. I've found many machines with 4 gigs of ram to not be enough to get decent performance from a database these days. I would suggest going with AMD and getting a board that can do at least 8GB if not 16 or even 32GB of ram. Even with what I've been hearing in this thread about a 20% speed difference with Intel parts, you will totally be ruined once you hit the 4GB barrier on your Intel hardware. -Alfred * Palle Girgensohn [070910 03:16] wrote: > Hi, > > We are looking at getting a server for running postgresql. Only postgresql, > a dedicated machine. Since we know FreeBSD very well, we plan on using it > as the OS. > > We have an offer for an IBM server, x3650, with 2 * DC Intel Xeon 5160, > Raid with two clusters, one for database and one for xlog. Loads and loads > of RAM. > > Now, I hear rumors that AMD is to be preferred over Intel for performance > reasons. Any comments on that? Are we better off getting AMD processors, > when running Unix? > > Regards, > Palle > > _______________________________________________ > 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" -- - Alfred Perlstein From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 23:32:23 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 5E4E616A419; Mon, 10 Sep 2007 23:32:23 +0000 (UTC) (envelope-from kris@FreeBSD.org) Received: from weak.local (hub.freebsd.org [IPv6:2001:4f8:fff6::36]) by mx1.freebsd.org (Postfix) with ESMTP id 5933913C468; Mon, 10 Sep 2007 23:32:20 +0000 (UTC) (envelope-from kris@FreeBSD.org) Message-ID: <46E5D402.8060305@FreeBSD.org> Date: Tue, 11 Sep 2007 01:32:18 +0200 From: Kris Kennaway User-Agent: Thunderbird 2.0.0.6 (Macintosh/20070728) MIME-Version: 1.0 To: Alfred Perlstein References: <20070910224503.GO79417@elvis.mu.org> In-Reply-To: <20070910224503.GO79417@elvis.mu.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org, Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 23:32:23 -0000 Alfred Perlstein wrote: > Palle, > > I really haven't kept pace with Intel versus AMD in a while, my > understanding is that AMD is still the only 64bit game in town. > > For a database, the more memory you can get, the better. > > I've found many machines with 4 gigs of ram to not be enough to > get decent performance from a database these days. > > I would suggest going with AMD and getting a board that can > do at least 8GB if not 16 or even 32GB of ram. > > Even with what I've been hearing in this thread about a 20% speed > difference with Intel parts, you will totally be ruined once you > hit the 4GB barrier on your Intel hardware. That's actually not true, intel came out with their first amd64 clone (which they call "EM64T") something like 3 or 4 years ago. I cannot say from first hand experience but I have heard that their current generation is solidly outperforming amd64. Kris From owner-freebsd-performance@FreeBSD.ORG Mon Sep 10 23:57:13 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 567E416A419 for ; Mon, 10 Sep 2007 23:57:13 +0000 (UTC) (envelope-from paul@pathiakis.com) Received: from que02.charter.net (que02.charter.net [209.225.8.190]) by mx1.freebsd.org (Postfix) with ESMTP id 0239913C428 for ; Mon, 10 Sep 2007 23:57:12 +0000 (UTC) (envelope-from paul@pathiakis.com) Received: from aa04.charter.net ([10.20.200.156]) by mtai03.charter.net (InterMail vM.7.08.02.00 201-2186-121-20061213) with ESMTP id <20070910231126.OOWU21729.mtai03.charter.net@aa04.charter.net>; Mon, 10 Sep 2007 19:11:26 -0400 Received: from pc4.atlantisservices.com ([66.189.25.162]) by aa04.charter.net with ESMTP id <20070910231126.EFLN1254.aa04.charter.net@pc4.atlantisservices.com>; Mon, 10 Sep 2007 19:11:26 -0400 From: Paul Pathiakis Organization: Myself To: freebsd-performance@freebsd.org Date: Mon, 10 Sep 2007 19:11:27 -0400 User-Agent: KMail/1.9.7 References: <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> In-Reply-To: <20070910184621.GA77744@cons.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200709101911.28469.paul@pathiakis.com> X-Chzlrs: 0 Cc: Martin Cracauer , Erich Dollansky , Palle Girgensohn , Paul Pathiakis Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Sep 2007 23:57:13 -0000 On Monday 10 September 2007 14:46:21 Martin Cracauer wrote: > For integer workloads Intel's Core2-base Xeons outperforms K8 (the > old-school AMD64) by about 25-30% per clock per core. K10 seems to be > 5-15% faster than K8 for integer workloads (I hope to run my benchmark > suite on one thi week or weekend). > > However, tasks that use multiple cores and have threads on cores > communicate a lot see both AMD architectures close the gap. > > Paul Pathiakis wrote on Mon, Sep 10, 2007 at 10:17:40AM -0400: > > Be very, very careful in purchasing Core 2 Duo. There are major > > problems with the chip that have been documented across the board. > > These have been blown out of proportion by Theo. Can you point to a > demonstratable case with current Linux or BSD kernels? > Agreed. However, Matt Dillon also made statements as did a few EE types. The chip is complicated due to poor design and the need for backward compatibility. I believe several people over the years have said that if they dumped everything pre-Pentium (486 instructions and earlier), the instruction set and complexity could easily be halved. Honestly, could you imagine how energy efficient and fast these chips (from both) would be at that point? One of the things that I'm seeing that really is starting to show is the use of more layers of cache and their increases in size. This is the same stop gap method everyone uses when they've hit a wall. P. From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 00:57:45 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 253CE16A417 for ; Tue, 11 Sep 2007 00:57:45 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: from smtpgate1.pacific.net.sg (smtpgate1.pacific.net.sg [203.120.90.31]) by mx1.freebsd.org (Postfix) with SMTP id 5DD3413C46E for ; Tue, 11 Sep 2007 00:57:44 +0000 (UTC) (envelope-from oceanare@pacific.net.sg) Received: (qmail 341 invoked from network); 11 Sep 2007 00:57:42 -0000 Received: from bb121-7-106-120.singnet.com.sg (HELO P2120.somewherefaraway.com) (oceanare@121.7.106.120) by smtpgate1.pacific.net.sg with ESMTPA; 11 Sep 2007 00:57:41 -0000 Message-ID: <46E5E7FC.4060306@pacific.net.sg> Date: Tue, 11 Sep 2007 08:57:32 +0800 From: Erich Dollansky User-Agent: Thunderbird 2.0.0.6 (X11/20070826) MIME-Version: 1.0 To: Martin Cracauer References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> In-Reply-To: <20070910184621.GA77744@cons.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable Cc: freebsd-performance@freebsd.org, Palle Girgensohn , Paul Pathiakis Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 00:57:45 -0000 Hi, Martin Cracauer wrote: > For integer workloads Intel's Core2-base Xeons outperforms K8 (the > old-school AMD64) by about 25-30% per clock per core. K10 seems to be > 5-15% faster than K8 for integer workloads (I hope to run my benchmark > suite on one thi week or weekend). >=20 the guys at heise.de published these numbers: Prozessoren Kerne Takt- SPEC CPU2006 (Base) im System frequenz Einzelkern Durchsatz int_2006 fp_2006 int_rate_2006 fp_rate_2006 2 =D7 Opteron 2350 2=D74 2,0 GHz 10,2 11,6 70,2 68,3 2 =D7 Opteron 2222 2=D72 3,0 GHz 12,8 13,9 50,0 50,2 2 =D7 Opteron 2212 2=D72 2,0 GHz (9,8) (10,4) (37,9) (38,6) 2 =D7 Xeon X5365 2=D74 3,0 GHz 15,7 (18,9) 15,8 72,5 (98,9) 61,9 Yes, the table looks like crap but you should be able to get the idea. The Xeons are faster if one CPU is running and slower when all are=20 running most likelz caused by the memory interface. The hardware was based on a Broadcom chipset and supported by SuSE but=20 not by other Linux versions. > On the other hand, if you want K8 or K10 in a modern SMP mainboard you > have to live with NVidia for chipsets, and the socket F boards all It seems that Broadcom is back into this game. Erich From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 06:11:59 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 4401A16A417 for ; Tue, 11 Sep 2007 06:11:59 +0000 (UTC) (envelope-from kometen@gmail.com) Received: from nf-out-0910.google.com (nf-out-0910.google.com [64.233.182.185]) by mx1.freebsd.org (Postfix) with ESMTP id D78AB13C457 for ; Tue, 11 Sep 2007 06:11:58 +0000 (UTC) (envelope-from kometen@gmail.com) Received: by nf-out-0910.google.com with SMTP id k4so989256nfd for ; Mon, 10 Sep 2007 23:11:57 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=sRj7plPdufK9TnqAJYXQjpSy1BJrpJXhtACJidmJy/U=; b=hpdi6ntD9L+M6nlpTGsx/JFD2YKSiyZaE3rFb/G/KQsa1U+vULHJYWgnSzA3yKXzPb7oiSVeefY2BZcs5rYmWGey+Jp1V/oR5YqZtkcv+oJufGkj1T4SzbkSQfTahl+ut/zuzCpe1I3UQP64h37ceZrjyaxsSj0HnefKGAaazDU= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=PHHg3PJojpav+XgDApYGQVSJ8cn70XCJTfmeoZu7mUbNjnsDRVGvZufdtXNvatnGdCbJhgbVplTSnhfbIoSI8NAmoTKOncTRRxqkiaiPM6kThYaXXtndQlD+DasHcyV7MtC473oGBAt0rSGIlndw9cR1FUiPbP/FtMYwv+0+O+A= Received: by 10.78.180.16 with SMTP id c16mr2337744huf.1189491117286; Mon, 10 Sep 2007 23:11:57 -0700 (PDT) Received: by 10.78.159.13 with HTTP; Mon, 10 Sep 2007 23:11:57 -0700 (PDT) Message-ID: Date: Tue, 11 Sep 2007 08:11:57 +0200 From: "Claus Guttesen" To: "Alfred Perlstein" In-Reply-To: <20070910224503.GO79417@elvis.mu.org> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <20070910224503.GO79417@elvis.mu.org> Cc: freebsd-performance@freebsd.org, Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 06:11:59 -0000 > Palle, > > I really haven't kept pace with Intel versus AMD in a while, my > understanding is that AMD is still the only 64bit game in town. > Pls. don't top post!! > > -Alfred > > * Palle Girgensohn [070910 03:16] wrote: > > Hi, > > > > We are looking at getting a server for running postgresql. Only postgresql, > > a dedicated machine. Since we know FreeBSD very well, we plan on using it > > as the OS. > > > > We have an offer for an IBM server, x3650, with 2 * DC Intel Xeon 5160, > > Raid with two clusters, one for database and one for xlog. Loads and loads > > of RAM. > > > > Now, I hear rumors that AMD is to be preferred over Intel for performance > > reasons. Any comments on that? Are we better off getting AMD processors, > > when running Unix? > > > > Regards, > > Palle > > > > _______________________________________________ > > 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" > > -- > - Alfred Perlstein > _______________________________________________ > 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" > -- regards Claus When lenity and cruelty play for a kingdom, the gentlest gamester is the soonest winner. Shakespeare From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 08:40:15 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 67EFD16A420 for ; Tue, 11 Sep 2007 08:40:15 +0000 (UTC) (envelope-from gofp-freebsd-performance@m.gmane.org) Received: from ciao.gmane.org (main.gmane.org [80.91.229.2]) by mx1.freebsd.org (Postfix) with ESMTP id 2DB8313C459 for ; Tue, 11 Sep 2007 08:40:15 +0000 (UTC) (envelope-from gofp-freebsd-performance@m.gmane.org) Received: from root by ciao.gmane.org with local (Exim 4.43) id 1IV0yY-0008Bo-R0 for freebsd-performance@freebsd.org; Tue, 11 Sep 2007 10:20:02 +0200 Received: from lara.cc.fer.hr ([161.53.72.113]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Tue, 11 Sep 2007 10:20:02 +0200 Received: from ivoras by lara.cc.fer.hr with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Tue, 11 Sep 2007 10:20:02 +0200 X-Injected-Via-Gmane: http://gmane.org/ To: freebsd-performance@freebsd.org From: Ivan Voras Date: Mon, 10 Sep 2007 15:26:59 +0200 Lines: 34 Message-ID: References: Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-ripemd160; protocol="application/pgp-signature"; boundary="------------enig982D0587DCE8BF028197D9DF" X-Complaints-To: usenet@sea.gmane.org X-Gmane-NNTP-Posting-Host: lara.cc.fer.hr User-Agent: Thunderbird 1.5.0.12 (X11/20060911) In-Reply-To: X-Enigmail-Version: 0.94.4.0 Sender: news X-UiO-SPF-Received: Received-SPF: pass (mail-mx2.uio.no: domain of sea.gmane.org designates 80.91.229.5 as permitted sender) client-ip=80.91.229.5; envelope-from=news@sea.gmane.org; helo=sea.gmane.org; X-UiO-Spam-info: not spam, SpamAssassin (score=0.0, required=12.0, autolearn=disabled, none) X-UiO-Scanned: 94426EECFE2F5BD3467183F16A2B0922B3C8655F X-UiO-SPAM-Test: remote_host: 80.91.229.5 spam_score: 0 maxlevel 200 minaction 2 bait 0 mail/h: 18 total 370 max/h 30 blacklist 0 greylist 0 ratelimit 0 Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 08:40:15 -0000 This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enig982D0587DCE8BF028197D9DF Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: quoted-printable Palle Girgensohn wrote: > Hi, >=20 > We are looking at getting a server for running postgresql. Only=20 > postgresql, a dedicated machine. Since we know FreeBSD very well, we=20 > plan on using it as the OS. You might want to wait a little until 7.0 or until some more important=20 bits get MFC'ed to 6.x: http://blog.insidesystems.net/articles/2007/04/11/postgresql-scaling-on-6= -2-and-7-0 --------------enig982D0587DCE8BF028197D9DF Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFG5UYrldnAQVacBcgRA+XiAKCnlfGYD8FqUbaJmz+0zGmxJNfL9gCdFRr5 TCPpv4NdWl8raKC6Rgh2XHA= =tNSo -----END PGP SIGNATURE----- --------------enig982D0587DCE8BF028197D9DF-- From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 09:18:14 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B0AF416A418 for ; Tue, 11 Sep 2007 09:18:14 +0000 (UTC) (envelope-from almarrie@gmail.com) Received: from nf-out-0910.google.com (nf-out-0910.google.com [64.233.182.187]) by mx1.freebsd.org (Postfix) with ESMTP id 58DD413C45A for ; Tue, 11 Sep 2007 09:18:14 +0000 (UTC) (envelope-from almarrie@gmail.com) Received: by nf-out-0910.google.com with SMTP id k4so1019705nfd for ; Tue, 11 Sep 2007 02:18:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=sK9CP2ENaDl7JlI6u0QTYK5Ej/0/pFTTnN44uZ2CC0c=; b=mGFvtR2m/E6jo6jL1CDtN1flW6m6YSXae0riPtJ04iqxydlk0520gdXW5tCOmR4cdKtaql8nRq07bEZyKNVE32S8N86THdZBb/RTJUfadKKlr8Z/jkhb5vhfWVE6qNkJov9vSvTfvxu8+udGvE7EA6Rlz74Ku59L2gFjEfv4as8= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=iHCsXAoMowMxGaTBy+cWq47mc5jN+lpRu9V+dkLyB7hpSM7sOKMV7cGgLju+ycJ/wtT41qAhSDikW9ZguM8GUDZ03fl2ooivSA0zw9MeEm0eMffhJ7bMO3Na79QkYW7KEq66IpNr7x4Ali6yMHETCBCGH7/Z50yT9AhtQDvhp7g= Received: by 10.86.25.17 with SMTP id 17mr4444069fgy.1189500665513; Tue, 11 Sep 2007 01:51:05 -0700 (PDT) Received: by 10.86.2.1 with HTTP; Tue, 11 Sep 2007 01:51:05 -0700 (PDT) Message-ID: <499c70c0709110151i3ce68598v2f6bc5d9bce4b6c9@mail.gmail.com> Date: Tue, 11 Sep 2007 11:51:05 +0300 From: "Abdullah Ibn Hamad Al-Marri" To: freebsd-performance@freebsd.org In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 09:18:14 -0000 On 9/10/07, Ivan Voras wrote: > Palle Girgensohn wrote: > > Hi, > > > > We are looking at getting a server for running postgresql. Only > > postgresql, a dedicated machine. Since we know FreeBSD very well, we > > plan on using it as the OS. > > You might want to wait a little until 7.0 or until some more important > bits get MFC'ed to 6.x: > > http://blog.insidesystems.net/articles/2007/04/11/postgresql-scaling-on-6-2-and-7-0 I second this, FreeBSD 7.0-CURRENT is frozen so I guess it's ok to use it, and I use it my production server which really enjoy the C2D E660 and 2 GB ram, with AMD64 arch. -- Regards, -Abdullah Ibn Hamad Al-Marri Arab Portal http://www.WeArab.Net/ From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 11:43:19 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 181D916A418 for ; Tue, 11 Sep 2007 11:43:19 +0000 (UTC) (envelope-from borjamar@sarenet.es) Received: from proxypop1.sarenet.es (proxypop1.sarenet.es [194.30.0.99]) by mx1.freebsd.org (Postfix) with ESMTP id 852F813C46C for ; Tue, 11 Sep 2007 11:43:18 +0000 (UTC) (envelope-from borjamar@sarenet.es) Received: from [127.0.0.1] (unknown [212.81.200.214]) by proxypop1.sarenet.es (Postfix) with ESMTP id 8933D5CE5 for ; Tue, 11 Sep 2007 13:43:16 +0200 (CEST) Mime-Version: 1.0 (Apple Message framework v752.2) Content-Transfer-Encoding: 7bit Message-Id: Content-Type: text/plain; charset=US-ASCII; format=flowed To: freebsd-performance@freebsd.org From: Borja Marcos Date: Tue, 11 Sep 2007 13:43:13 +0200 X-Mailer: Apple Mail (2.752.2) Subject: Announcement: Devilator 1.0a for FreeBSD X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 11:43:19 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, I'm happy to announce the first public release (consider it an alpha) of Devilator, a performance data collector for FreeBSD systems that generates Orca compatible files. The README file follows. ************************************************************************ ******* WARNING!!!! This is the first public evaluation release of a program developed for internal use. Please exercise extreme care when using it, and report any issues to the author to have them fixed in future releases. ************************************************************************ ******* ================================== DEVILATOR ================================== "Why does my system run slowly?" "What can I do to make it run faster?" "The CPU seems to be idle. Why don't I get better performance?" "Will a bigger CPU make it run faster?" "Will I need a better CPU next year?" "Was the system slower last Thursday, or is it just my imagination?" These questions are made probably hundreds or perhaps even thousands of times every day, and the answer is not always easy. A system's performance is a function of its CPU power, bus bandwidth, memory capacity, disk I/O bandwidth (often neglected) and software. Nowadays, with the usage of concurrent programs growing, software performance problems get increasingly important as well. We can have the fastest and greatest hardware available, but a poor locking decision in a concurrent program can make it crawl, no matter what we do. A bad mix of workloads can cause these problems as well. Performance tuning is often a black art, and historical data is reall valuable, a need that the classical commands such as top, vmstat, gstat, etc don't address adequately (in my opinion). Devilator is a system performance data collector for FreeBSD. It's designed to work with Blair Zajac's Orca system (http://www.orcaware.com/ orca/) and it aims to be a FreeBSD equivalent of the "Orcallator" data collector for Solaris. This program has been growing for a long time, supplying performance data for our in-house systems. As this program has been written as a series of quick hacks, it is actually quite ugly, but effective anyway. The main goal when writing it was to gather enough performance data from the OS, always avoiding typical scripts that launch many other utilities. This program reads OS data directly, either through the sysctl interface or from GEOM (in the case of disk I/O performance data). As the documentation on some of these issues is not great, I have borrowed code from the GEOM utility "gstat", from "top", and "orcallator.se", part of the Orca package. The data collected by Devilator can be divided among the following groups: 1- Processor/s usage. This includes the classical "system load" values, and the percentage of CPU time spent on user processes, system, interrupt and nice. 2- OS activity. Includes a graph of forks/s, system calls/s, system processes and interrupt activity. The system processes and interrupt information show where the system is spending its share of CPU time. The system processes shown vary from version to version, with more information offered by the most recent FreeBSD releases. Hopefully this can be useful for system administrators and system developers alike, keeping historic data on the activity of tasks like "vmdaemon", "syncer", etc. 3- Process activity. Includes the number of processes in the system, how many of them are in a given state (running, sleeping, stopped, zombies, waiting or locked), and a graph of sleep situations. How many processes are sleeping on select(), waiting for network I/O, waiting for the disk subsystem, filesystem, etc? This is really useful when evaluating the performance of programs such as MySQL, trying different filesystem configurations or threading models. These data will be hopefully useful to FreeBSD developers and porters as well. 4- Network activity, offering the classical information on established connections, TCP and UDP traffic, network interface traffic and errors, TCP connection rate, etc. 5- Virtual memory, including swap and paging activity, memory usage, and page scanner activity. 6- Disk I/O data (available on FreeBSD 5+, as it depends on GEOM), including the percentage of busy time for disks, partitions and slices, together with the volume of data transfers in MB/s and the average service time in ms. - --------------- USING DEVILATOR - --------------- In order to use this program it's a good idea to be familiarised with Orca first. Orca (http://www.orcaware.com/orca/) is a general purpose graphing program based on RRDtool. This program can be used to monitor a single machine, running both Devilator and Orca on it, or it can be installed at dozens of servers, where it will just gather performance data and write it to text files, while a different server downloads those performance data files periodically, running Orca and generating the HTML files and RRD graphs for all of them. Depending on the number of machines, a single machine can monitor dozens of servers, and the impact of Devilator is negligible. It just wakes up at 5 minute intervals, reads some OS statistics using sysctl(3) and GEOM, and writes a line to a text file. In order to have Orca use the data provided by Devilator, the supplied Orca configuration file, devilator.cfg, must be used. The relevant variables to set up are: - -------------- # base_dir is the directory where the RRD files will be stored. base_dir /var/orca/rrd/orcallator # html_dir is the directory where the web pages (HTML files and PNG graphs) # will be stored. This directory will be made available with a suitable # web server such as Apache. html_dir /var/www/orca # find_files specifies the directory under which the text files with # performance data reside. It is possible to monitor a set of machines with a # single Orca instance, with each machine having its own directory under # /var/orca/orcallator, such as /var/orca/orcallator/server1, # /var/orca/orcallator/server2... find_files /var/orca/orcallator/(.*)/(?:(?:orcallator)| (?:percol))- \d{4}-\d{2}-\d{2}(?:-\d{3,})?(?:\.(?:Z|gz|bz2))? - --------------- Please read the Orca documentation and become familiar with it. - ------------------- FUTURE WORK - ------------------- The graphs are a bit messy right now, although they are easy to customize editing the .cfg file. This program generates a lot of information that might not be necessary for everyone. Unnecessary graphs can be turned off just by commenting out the relevant sections in the configuration file. Some enhancements would probably be useful. Tracking resources usage by process name would allow the administrator to know which CGI scripts are using more resources, etc. I will be happy to hear about other interesting datasets. Simple sysctl fetches are trivial to do. But obviously it would be insane to include all the available data. - -------------------- PORTABILITY - -------------------- This program has been developed under FreeBSD, and some of the bits (like its reliance on GEOM to gather disk I/O statistics) are pretty FreeBSD-specific. So far I have tested it and used it on FreeBSD 4.9 (x86), FreeBSD 6.x on sparc64 and x86, and FreeBSD 7 on amd64. - -------------------- SUPPORT - -------------------- This program is supplied as is, and there's absolutely no guarantee of any kind, including suitability for any particular purpose. There's no guarantee of technical support either. This program started its life as an internal development, and I'm polishing it for public distribution. Most of the development time has been payed for by my employer, Sarenet, and it's being released as a contribution to the FreeBSD community. Please send bugs, ideas, flames, etc to the following address: Borja Marcos. Sarenet S.A. borjamar@sarenet.es borjam@gmail.com (Preferred for Devilator issues) Devilator can be downloaded from: http://homepage.mac.com/borjam/.Public/devilator-1.0a.tar.gz MD5 (devilator-1.0a.tar.gz) = 5868761c2ce0f8dc3ff51167fd06400a And the PGP signature for the file follows: - -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.5.3 (Build 5003) qANQR1DEDQMAAhGlPtFtULJLjAHCPwMFAEbmbwOlPtFtULJLjBECYlEAnj1HejG2 rBJVYur8VvZ51MlSYVX1AJ9Aak05xRRkQ/BKn8Asl9nWsxD6rA== =rKMg - -----END PGP SIGNATURE----- - ---------------- "The thing he realised about the windows was this: because they had been converted into openable windows after they had first been designed to be impregnable, they were, in fact, much less secure than if they had been designed as openable windows in the first place." Douglas Adams, "Mostly Harmless" -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.5.3 (Build 5003) wj8DBQFG5nfqpT7RbVCyS4wRAr7eAJ98PUckfBpE6nH+ouU9hpN/FhT0LACg03UH ETLJnR4KJFgk02lNuTZanW4= =ULQE -----END PGP SIGNATURE----- From owner-freebsd-performance@FreeBSD.ORG Tue Sep 11 14:34:51 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id CC73116A417 for ; Tue, 11 Sep 2007 14:34:51 +0000 (UTC) (envelope-from cracauer@koef.zs64.net) Received: from koef.zs64.net (koef.zs64.net [212.12.50.230]) by mx1.freebsd.org (Postfix) with ESMTP id 602C013C46A for ; Tue, 11 Sep 2007 14:34:50 +0000 (UTC) (envelope-from cracauer@koef.zs64.net) Received: from koef.zs64.net (koef.zs64.net [212.12.50.230]) by koef.zs64.net (8.14.1/8.14.1) with ESMTP id l8BEYnaE079746; Tue, 11 Sep 2007 16:34:49 +0200 (CEST) (envelope-from cracauer@koef.zs64.net) Received: (from cracauer@localhost) by koef.zs64.net (8.14.1/8.14.1/Submit) id l8BEYmkl079745; Tue, 11 Sep 2007 10:34:48 -0400 (EDT) (envelope-from cracauer) Date: Tue, 11 Sep 2007 10:34:48 -0400 From: Martin Cracauer To: Paul Pathiakis Message-ID: <20070911143448.GA79526@cons.org> References: <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> <200709101911.28469.paul@pathiakis.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200709101911.28469.paul@pathiakis.com> User-Agent: Mutt/1.4.2.3i Cc: Erich Dollansky , freebsd-performance@freebsd.org, Martin Cracauer , Palle Girgensohn , Paul Pathiakis Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 11 Sep 2007 14:34:51 -0000 Paul Pathiakis wrote on Mon, Sep 10, 2007 at 07:11:27PM -0400: > On Monday 10 September 2007 14:46:21 Martin Cracauer wrote: > > For integer workloads Intel's Core2-base Xeons outperforms K8 (the > > old-school AMD64) by about 25-30% per clock per core. K10 seems to be > > 5-15% faster than K8 for integer workloads (I hope to run my benchmark > > suite on one thi week or weekend). > > > > However, tasks that use multiple cores and have threads on cores > > communicate a lot see both AMD architectures close the gap. > > > > Paul Pathiakis wrote on Mon, Sep 10, 2007 at 10:17:40AM -0400: > > > Be very, very careful in purchasing Core 2 Duo. There are major > > > problems with the chip that have been documented across the board. > > > > These have been blown out of proportion by Theo. Can you point to a > > demonstratable case with current Linux or BSD kernels? > > > Agreed. However, Matt Dillon also made statements as did a few EE types. > > The chip is complicated due to poor design and the need for backward > compatibility. I believe several people over the years have said that if > they dumped everything pre-Pentium (486 instructions and earlier), the > instruction set and complexity could easily be halved. Doubtful. All the legacy instructions that are not part of the "useful" set are just high-level microcode programs. You can tell by how slow they are :-) The true complexity of Core2 is in the new caches, including speculative prefetch (aka they keep a dependency graph around and can invalidate wrongly fetched cache lines). Most of the bugs that Theo was concerned about are in emmory management and MMU, which might or might not be made worse by the caches. There probably is some additional memory management complexity from i386 compatibility, but I don't see how, for example, the need to run non-MMU code would cause MMU bugs. Also, I really like to continue to use my bootloaders :-) > Honestly, could you imagine how energy efficient and fast these chips (from > both) would be at that point? > > One of the things that I'm seeing that really is starting to show is the use > of more layers of cache and their increases in size. Not sure what you mean here. The L2 and L3 caches we see today have been around for long. Martin -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Martin Cracauer http://www.cons.org/cracauer/ FreeBSD - where you want to go, today. http://www.freebsd.org/ From owner-freebsd-performance@FreeBSD.ORG Wed Sep 12 23:44:44 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 36E9E16A417 for ; Wed, 12 Sep 2007 23:44:44 +0000 (UTC) (envelope-from obrien@NUXI.org) Received: from dragon.nuxi.org (trang.nuxi.org [74.95.12.85]) by mx1.freebsd.org (Postfix) with ESMTP id BB08213C4CE for ; Wed, 12 Sep 2007 23:44:42 +0000 (UTC) (envelope-from obrien@NUXI.org) Received: from dragon.nuxi.org (obrien@localhost [127.0.0.1]) by dragon.nuxi.org (8.14.1/8.14.1) with ESMTP id l8CN8o2g084471; Wed, 12 Sep 2007 16:08:50 -0700 (PDT) (envelope-from obrien@dragon.nuxi.org) Received: (from obrien@localhost) by dragon.nuxi.org (8.14.1/8.14.1/Submit) id l8CN8mAr084469; Wed, 12 Sep 2007 16:08:48 -0700 (PDT) (envelope-from obrien) Date: Wed, 12 Sep 2007 16:08:48 -0700 From: "David O'Brien" To: Erich Dollansky Message-ID: <20070912230848.GB75569@dragon.NUXI.org> References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> <46E5E7FC.4060306@pacific.net.sg> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <46E5E7FC.4060306@pacific.net.sg> X-Operating-System: FreeBSD 7.0-CURRENT Organization: The NUXI BSD Group X-Pgp-Rsa-Fingerprint: B7 4D 3E E9 11 39 5F A3 90 76 5D 69 58 D9 98 7A X-Pgp-Rsa-Keyid: 1024/34F9F9D5 User-Agent: Mutt/1.5.11 Cc: Martin Cracauer , freebsd-performance@freebsd.org, Palle Girgensohn , Paul Pathiakis Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: obrien@freebsd.org List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 12 Sep 2007 23:44:44 -0000 On Tue, Sep 11, 2007 at 08:57:32AM +0800, Erich Dollansky wrote: > It seems that Broadcom is back into this game. So far the Broadcom Serverworks HT1000 SATA controller is POO. I'm seeing all kinds of disk corruption with FreeBSD on a Tyan s3992. Googling shows that other FreeBSD users have been less than sucessful with running on the HT1000. -- -- David (obrien@FreeBSD.org) Q: Because it reverses the logical flow of conversation. A: Why is top-posting (putting a reply at the top of the message) frowned upon? Let's not play "Jeopardy-style quoting" From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 03:28:01 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 4A09E16A417; Thu, 13 Sep 2007 03:28:01 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 4807413C428; Thu, 13 Sep 2007 03:28:01 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: by elvis.mu.org (Postfix, from userid 1192) id EC9541A4D7C; Wed, 12 Sep 2007 20:28:00 -0700 (PDT) Date: Wed, 12 Sep 2007 20:28:00 -0700 From: Alfred Perlstein To: Kris Kennaway Message-ID: <20070913032800.GO79417@elvis.mu.org> References: <20070910224503.GO79417@elvis.mu.org> <46E5D402.8060305@FreeBSD.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <46E5D402.8060305@FreeBSD.org> User-Agent: Mutt/1.4.2.3i Cc: freebsd-performance@freebsd.org, Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 03:28:01 -0000 * Kris Kennaway [070910 16:32] wrote: > Alfred Perlstein wrote: > >Palle, > > > >I really haven't kept pace with Intel versus AMD in a while, my > >understanding is that AMD is still the only 64bit game in town. > > > >For a database, the more memory you can get, the better. > > > >I've found many machines with 4 gigs of ram to not be enough to > >get decent performance from a database these days. > > > >I would suggest going with AMD and getting a board that can > >do at least 8GB if not 16 or even 32GB of ram. > > > >Even with what I've been hearing in this thread about a 20% speed > >difference with Intel parts, you will totally be ruined once you > >hit the 4GB barrier on your Intel hardware. > > That's actually not true, intel came out with their first amd64 clone > (which they call "EM64T") something like 3 or 4 years ago. I cannot say > from first hand experience but I have heard that their current > generation is solidly outperforming amd64. Actually, what I said was true, it was my understanding that was wrong. :) I guess the answer I was trying to say was, go for whatever gives you room for a lot of RAM. -- - Alfred Perlstein From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 06:49:57 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 8F0A316A41A; Thu, 13 Sep 2007 06:49:57 +0000 (UTC) (envelope-from kris@FreeBSD.org) Received: from weak.local (hub.freebsd.org [IPv6:2001:4f8:fff6::36]) by mx1.freebsd.org (Postfix) with ESMTP id C00BC13C458; Thu, 13 Sep 2007 06:49:52 +0000 (UTC) (envelope-from kris@FreeBSD.org) Message-ID: <46E8DD8E.8070706@FreeBSD.org> Date: Thu, 13 Sep 2007 08:49:50 +0200 From: Kris Kennaway User-Agent: Thunderbird 2.0.0.6 (Macintosh/20070728) MIME-Version: 1.0 To: Alfred Perlstein References: <20070910224503.GO79417@elvis.mu.org> <46E5D402.8060305@FreeBSD.org> <20070913032800.GO79417@elvis.mu.org> In-Reply-To: <20070913032800.GO79417@elvis.mu.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org, Palle Girgensohn Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 06:49:57 -0000 Alfred Perlstein wrote: > * Kris Kennaway [070910 16:32] wrote: >> Alfred Perlstein wrote: >>> Palle, >>> >>> I really haven't kept pace with Intel versus AMD in a while, my >>> understanding is that AMD is still the only 64bit game in town. >>> >>> For a database, the more memory you can get, the better. >>> >>> I've found many machines with 4 gigs of ram to not be enough to >>> get decent performance from a database these days. >>> >>> I would suggest going with AMD and getting a board that can >>> do at least 8GB if not 16 or even 32GB of ram. >>> >>> Even with what I've been hearing in this thread about a 20% speed >>> difference with Intel parts, you will totally be ruined once you >>> hit the 4GB barrier on your Intel hardware. >> That's actually not true, intel came out with their first amd64 clone >> (which they call "EM64T") something like 3 or 4 years ago. I cannot say >> from first hand experience but I have heard that their current >> generation is solidly outperforming amd64. > > Actually, what I said was true, it was my understanding that was > wrong. :) > > I guess the answer I was trying to say was, go for whatever > gives you room for a lot of RAM. > Yep, that is still good advice. Kris From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 13:43:17 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 238A216A41B for ; Thu, 13 Sep 2007 13:43:17 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from melon.pingpong.net (melon.pingpong.net [195.178.174.161]) by mx1.freebsd.org (Postfix) with ESMTP id B73E913C46C for ; Thu, 13 Sep 2007 13:43:16 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from localhost (localhost.pingpong.net [127.0.0.1]) by melon.pingpong.net (Postfix) with ESMTP id 273495087C; Thu, 13 Sep 2007 15:43:14 +0200 (CEST) Received: from melon.pingpong.net ([127.0.0.1]) by localhost (melon.pingpong.net [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 13247-01-4; Thu, 13 Sep 2007 15:43:13 +0200 (CEST) Received: from [192.168.1.187] (unknown [213.136.40.204]) by melon.pingpong.net (Postfix) with ESMTP id C1FF65084D; Thu, 13 Sep 2007 15:43:12 +0200 (CEST) Date: Thu, 13 Sep 2007 15:43:12 +0200 From: Palle Girgensohn To: obrien@freebsd.org, Erich Dollansky Message-ID: In-Reply-To: <20070912230848.GB75569@dragon.NUXI.org> References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> <46E5E7FC.4060306@pacific.net.sg> <20070912230848.GB75569@dragon.NUXI.org> X-Mailer: Mulberry/4.0.8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-Virus-Scanned: by amavisd-new at pingpong.net Cc: Martin Cracauer , freebsd-performance@freebsd.org, Paul Pathiakis Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 13:43:17 -0000 --On onsdag, september 12, 2007 16.08.48 -0700 David O'Brien wrote: > On Tue, Sep 11, 2007 at 08:57:32AM +0800, Erich Dollansky wrote: >> It seems that Broadcom is back into this game. > > So far the Broadcom Serverworks HT1000 SATA controller is POO. > I'm seeing all kinds of disk corruption with FreeBSD on a Tyan s3992. > Googling shows that other FreeBSD users have been less than sucessful > with running on the HT1000. Hmm... What controller does HP use for their ProLiant DL380? "HP Smart Array P400" is the RAID I can expect. googling implies that Broadcom is involved? From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 18:16:24 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id ED8CA16A419 for ; Thu, 13 Sep 2007 18:16:24 +0000 (UTC) (envelope-from kometen@gmail.com) Received: from nf-out-0910.google.com (nf-out-0910.google.com [64.233.182.189]) by mx1.freebsd.org (Postfix) with ESMTP id 6DF5F13C458 for ; Thu, 13 Sep 2007 18:16:24 +0000 (UTC) (envelope-from kometen@gmail.com) Received: by nf-out-0910.google.com with SMTP id b2so557005nfb for ; Thu, 13 Sep 2007 11:16:23 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=rBsh/IwfboJIJ6cgs5nXQd6wspTJ1XTXGUD7OjAZT24=; b=cFGHAyAVMcCuFMrhe/fvkONgnjfXsvhf4OgKqu8T1KOlpg2n83TB1QK/886nJHzPHVvESsLDVYkEFH1mDCUu5yE5UUr2Ed6uV7QgrFGKelbq26Zjd4GN3OiN3pCpnnOF8h58doPJtHuP0rQW1WtYFX7uOQ6R8X4lFFoPSiPSt54= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=SLPWSaPya7DdUbkjLt4j+1EaDJ0LZqURPhsrJ25qFt2BY2opqqJc2txSd/60R2DBXoQHSepgjgzhwX9nMTgtvKqzQRioh8sI6by+I94l/5H3jWoxaTzh4eJjQhpFRl1E4y4yLPnSjmoSIr/ui9sdyZQwHzs4O9FixcONuXdaxy4= Received: by 10.78.170.17 with SMTP id s17mr545019hue.1189707382306; Thu, 13 Sep 2007 11:16:22 -0700 (PDT) Received: by 10.78.159.13 with HTTP; Thu, 13 Sep 2007 11:16:22 -0700 (PDT) Message-ID: Date: Thu, 13 Sep 2007 20:16:22 +0200 From: "Claus Guttesen" To: "Palle Girgensohn" In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <46E51D32.5070803@pacific.net.sg> <74611C396F3A6D027B0E8ABA@rambutan.pingpong.net> <46E52260.60000@pacific.net.sg> <46E55204.20905@eagleaccess.com> <20070910184621.GA77744@cons.org> <46E5E7FC.4060306@pacific.net.sg> <20070912230848.GB75569@dragon.NUXI.org> Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 18:16:25 -0000 > > On Tue, Sep 11, 2007 at 08:57:32AM +0800, Erich Dollansky wrote: > >> It seems that Broadcom is back into this game. > > > > So far the Broadcom Serverworks HT1000 SATA controller is POO. > > I'm seeing all kinds of disk corruption with FreeBSD on a Tyan s3992. > > Googling shows that other FreeBSD users have been less than sucessful > > with running on the HT1000. > > Hmm... What controller does HP use for their ProLiant DL380? "HP Smart > Array P400" is the RAID I can expect. googling implies that Broadcom is > involved? ciss. I have one DL380 G5 (woodcrest @ 3 GHz and 16 GB ram) running FreeBSD 6.2 RC1 running postgresql and one DL360 G5 (clowertown @ 2 GHz and 8 GB ram) running 7.0 doing web. But whether it's broadcom on the raid-controller I can't tell. It "just" works. -- regards Claus When lenity and cruelty play for a kingdom, the gentlest gamester is the soonest winner. Shakespeare From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 19:25:33 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 634A616A417 for ; Thu, 13 Sep 2007 19:25:33 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from zoraida.natserv.net (p65-147.acedsl.com [66.114.65.147]) by mx1.freebsd.org (Postfix) with ESMTP id 2B95013C45A for ; Thu, 13 Sep 2007 19:25:33 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from 35st.simplicato.com (static-71-249-233-130.nycmny.east.verizon.net [71.249.233.130]) by zoraida.natserv.net (Postfix) with ESMTP id 91E5A17036; Thu, 13 Sep 2007 15:08:45 -0400 (EDT) References: Message-ID: X-Mailer: http://www.courier-mta.org/cone/ From: Francisco Reyes To: Palle Girgensohn Date: Thu, 13 Sep 2007 15:07:17 -0400 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="US-ASCII" Content-Disposition: inline Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 19:25:33 -0000 Palle Girgensohn writes: > Now, I hear rumors that AMD is to be preferred over Intel for performance >From what I have read in the past, specially in the postgresql list, it seems the AMD64 cpus do better with Postgresql. Possibly because of better bus architecture. Personally I think having lots of memory and a good disk subsystem is going to make a bigger difference. What sizes are you expecting for your data? What type of growth? How many concurrent connections? From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 20:45:20 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 1724416A46C for ; Thu, 13 Sep 2007 20:45:20 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from proxy2.bredband.net (proxy2.bredband.net [195.54.101.72]) by mx1.freebsd.org (Postfix) with ESMTP id C1DA713C465 for ; Thu, 13 Sep 2007 20:45:19 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se (85.225.86.47) by proxy2.bredband.net (7.3.127) id 46E932370002108D; Thu, 13 Sep 2007 22:24:38 +0200 Date: Thu, 13 Sep 2007 22:24:38 +0200 From: Palle Girgensohn To: Francisco Reyes Message-ID: <1F219879A7E5C565C96109FF@c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se> In-Reply-To: References: X-Mailer: Mulberry/4.0.6 (Mac OS X) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 20:45:20 -0000 --On torsdag, torsdag 13 sep 2007 15.07.17 -0400 Francisco Reyes wrote: > Palle Girgensohn writes: > >> Now, I hear rumors that AMD is to be preferred over Intel for >> performance > > From what I have read in the past, specially in the postgresql list, it > seems the AMD64 cpus do better with Postgresql. Possibly because of > better bus architecture. I think this is not current information; the new woodcrest architecture performs mucg better, although this is deduced from this thread's discussion... > Personally I think having lots of memory and a good disk subsystem is > going to make a bigger difference. true > What sizes are you expecting for your data? > What type of growth? How many concurrent connections? Presently ~pgsql/data has a 16 GB footprint. The growth is rather slow, around a percent per week, sometimes much more due to batch updates. Even with a high system load, we seldom see more than a 30-40 concurrent postgresql connections. /Palle From owner-freebsd-performance@FreeBSD.ORG Thu Sep 13 23:44:51 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C565416A418 for ; Thu, 13 Sep 2007 23:44:51 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from zoraida.natserv.net (p65-147.acedsl.com [66.114.65.147]) by mx1.freebsd.org (Postfix) with ESMTP id 8DD3813C457 for ; Thu, 13 Sep 2007 23:44:51 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from 35st.simplicato.com (static-71-249-233-130.nycmny.east.verizon.net [71.249.233.130]) by zoraida.natserv.net (Postfix) with ESMTP id BE19B17036; Thu, 13 Sep 2007 19:46:19 -0400 (EDT) References: <1F219879A7E5C565C96109FF@c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se> Message-ID: X-Mailer: http://www.courier-mta.org/cone/ From: Francisco Reyes To: Palle Girgensohn Date: Thu, 13 Sep 2007 19:44:50 -0400 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="US-ASCII" Content-Disposition: inline Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 13 Sep 2007 23:44:51 -0000 Palle Girgensohn writes: > Presently ~pgsql/data has a 16 GB footprint. If you can put 4GB or better in your machine you should do well. Specially since you mentioned you are mostly read with relatively small amount of writes. > The growth is rather slow, around a percent per week What controller are you getting? We have a 3ware SATA controller with RAID6 and it performs pretty well. Based on what you wrote SATA RAID should be enough for your load and usage pattern. Obviously if you can afford SCSI/SAS performance will likely be even better. However make sure you can get management program for the controller. At the very least some type of notification if the raid is degraded. From owner-freebsd-performance@FreeBSD.ORG Fri Sep 14 00:21:07 2007 Return-Path: Delivered-To: performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 8B46316A417 for ; Fri, 14 Sep 2007 00:21:07 +0000 (UTC) (envelope-from jonny@jonny.eng.br) Received: from coe.ufrj.br (roma.coe.ufrj.br [146.164.53.65]) by mx1.freebsd.org (Postfix) with ESMTP id 3D38E13C45A for ; Fri, 14 Sep 2007 00:21:07 +0000 (UTC) (envelope-from jonny@jonny.eng.br) Received: from localhost (localhost [127.0.0.1]) by coe.ufrj.br (Postfix) with ESMTP id 41D52125418 for ; Thu, 13 Sep 2007 20:59:54 -0300 (BRT) X-Virus-Scanned: amavisd-new at coe.ufrj.br Received: from coe.ufrj.br ([146.164.53.65]) by localhost (roma.coe.ufrj.br [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IcMw2ZWn40Ra for ; Thu, 13 Sep 2007 20:59:48 -0300 (BRT) Received: from [201.19.103.168] (unknown [201.19.103.168]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by coe.ufrj.br (Postfix) with ESMTP id 313E9125409 for ; Thu, 13 Sep 2007 20:59:48 -0300 (BRT) Message-ID: <46E9CEF7.2070307@jonny.eng.br> Date: Thu, 13 Sep 2007 20:59:51 -0300 From: =?ISO-8859-1?Q?Jo=E3o_Carlos_Mendes_Lu=EDs?= User-Agent: Thunderbird 2.0.0.6 (Windows/20070728) MIME-Version: 1.0 To: performance@freebsd.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Cc: Subject: SATA mirrror performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 14 Sep 2007 00:21:07 -0000 Hi, I'm not sure if this is the right forum, but since the main keyword is performance, I'd try here first. I have just installed two 500G SATA discs from Seagate, model ST3500641AS in an ASUS M2N-E motherboard (nVidia MCP55 chipset). Since this is a home desktop and I need dual boot, I used nVidia's RAID technology to create the array, while still using the previous disks for booting. Well, the suggested device to control these disks, IIRC, is ataraid, so I went for it. But its performance was incredibly slow. I had less than half a megabyte per second in a raw transfer (dd). Formating 100G UFS2 partitions take more than a minute. And all this was drivers fault, as far as I could notice from "systat -v" output. The array was operating near 100% capacity. While running newfs, and this I remember for sure, the array was performing at 4 (four!) transfers per second, and near 100% load. I also noticed that ataraid does not integrate with GEOM. Shouldn't it be? Just to be sure it was no defect in disks, they worked perfectly in Windows XP. So my solution was to build a whole disk RAID1 device using gmirror, but now I have two independent and non-interoperational RAID technologies. Indeed, I am very luck that nVidia's RAID does not use the same sector as gmirror for metadata, or if they use, that it does not clash. After using gmirror, now I have the RAID in its full performance, getting over 60Mbytes per second at raw reads, very near the 70MBps from the specs. Could only be better if we already had NCQ working. Now the question: Is this expected? Is ataraid somehow deprecated? If that matters, this has been done on the last week's RELENG_6 source build. The CPU is an AMD Athlon 64 X2 5200+, with 3G RAM. Jonny -- João Carlos Mendes Luís - Networking Engineer - jonny@jonny.eng.br From owner-freebsd-performance@FreeBSD.ORG Fri Sep 14 06:44:51 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id A000D16A41B for ; Fri, 14 Sep 2007 06:44:51 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from proxy1.bredband.net (proxy1.bredband.net [195.54.101.71]) by mx1.freebsd.org (Postfix) with ESMTP id 6189313C468 for ; Fri, 14 Sep 2007 06:44:51 +0000 (UTC) (envelope-from girgen@pingpong.net) Received: from c-6254e155.1521-1-64736c12.cust.bredbandsbolaget.se (85.225.84.98) by proxy1.bredband.net (7.3.127) id 46E93261000403FB; Fri, 14 Sep 2007 08:44:47 +0200 Date: Fri, 14 Sep 2007 08:44:48 +0200 From: Palle Girgensohn To: Francisco Reyes Message-ID: <26F41A5DACB2D2CB43A5829E@c-6254e155.1521-1-64736c12.cust.bredbandsbolaget.se> In-Reply-To: References: <1F219879A7E5C565C96109FF@c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se> X-Mailer: Mulberry/4.0.6 (Mac OS X) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 14 Sep 2007 06:44:51 -0000 --On torsdag, torsdag 13 sep 2007 19.44.50 -0400 Francisco Reyes wrote: > Palle Girgensohn writes: > >> Presently ~pgsql/data has a 16 GB footprint. > > If you can put 4GB or better in your machine you should do well. > Specially since you mentioned you are mostly read with relatively small > amount of writes. > >> The growth is rather slow, around a percent per week Sorry, my mistake, more like a percent per day at the moment... We are planning about 16 GB RAM, actually. Maybe it is overkill? > What controller are you getting? > We have a 3ware SATA controller with RAID6 and it performs pretty well. > Based on what you wrote SATA RAID should be enough for your load and > usage pattern. > > Obviously if you can afford SCSI/SAS performance will likely be even > better. However make sure you can get management program for the > controller. At the very least some type of notification if the raid is > degraded. We will probably go for SCSI. HP DL380 with "HP SmartArray", aka ciss. From owner-freebsd-performance@FreeBSD.ORG Fri Sep 14 07:33:24 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B875516A46B for ; Fri, 14 Sep 2007 07:33:24 +0000 (UTC) (envelope-from kometen@gmail.com) Received: from nf-out-0910.google.com (nf-out-0910.google.com [64.233.182.184]) by mx1.freebsd.org (Postfix) with ESMTP id 4F2DE13C45D for ; Fri, 14 Sep 2007 07:33:24 +0000 (UTC) (envelope-from kometen@gmail.com) Received: by nf-out-0910.google.com with SMTP id b2so612001nfb for ; Fri, 14 Sep 2007 00:33:23 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=BqaY3nCW/G/9WLOcCEGbfo6DwgdJ85fx2k6kAPeu5bg=; b=WUblh8jGky3iT9GNv2WoRmYjYYEHc08Hq25uT+u9OZQfSPikttmB/E3ureu3P13uMuzh/OFXRgkDFqhYBN+FHHOPxFRrwGm/SJ5JqoAvcEEXdefkpIqH2SlWvIYq5PopAxBKcfJ4mVUN/8Ob5mt4tKRPVF2QeKnkYT9n34fm+Ho= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=rQJPFhh84cHMcsXOJPSbReoyVcHWieNnKIZqoXh76D1f6L8eW2MfTtVxaeART9vnJMtPg4ZomHUeRtX+Um04UkTfW3R+b08OuWbT3lel3VLiSY8Y1A5uLWkrLp7NTV9qnkX3KjqPHXq7NFCyhCeVC6uLPbg+tEX14o65+Bv21wI= Received: by 10.78.132.2 with SMTP id f2mr798012hud.1189755199525; Fri, 14 Sep 2007 00:33:19 -0700 (PDT) Received: by 10.78.159.13 with HTTP; Fri, 14 Sep 2007 00:33:19 -0700 (PDT) Message-ID: Date: Fri, 14 Sep 2007 09:33:19 +0200 From: "Claus Guttesen" To: "Palle Girgensohn" In-Reply-To: <26F41A5DACB2D2CB43A5829E@c-6254e155.1521-1-64736c12.cust.bredbandsbolaget.se> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <1F219879A7E5C565C96109FF@c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se> <26F41A5DACB2D2CB43A5829E@c-6254e155.1521-1-64736c12.cust.bredbandsbolaget.se> Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 14 Sep 2007 07:33:24 -0000 > > Obviously if you can afford SCSI/SAS performance will likely be even > > better. However make sure you can get management program for the > > controller. At the very least some type of notification if the raid is > > degraded. > > We will probably go for SCSI. HP DL380 with "HP SmartArray", aka ciss. This server is very solid and is fully supported by FreeBSD 6.2 and 7.0. We have one with 4 146 GB sas-disks in raid 1+0. Just make sure you get at least 256 MB cache. The raid-management tool is very easy to use. -- regards Claus When lenity and cruelty play for a kingdom, the gentlest gamester is the soonest winner. Shakespeare From owner-freebsd-performance@FreeBSD.ORG Fri Sep 14 22:44:38 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0A1DF16A419 for ; Fri, 14 Sep 2007 22:44:38 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from zoraida.natserv.net (p65-147.acedsl.com [66.114.65.147]) by mx1.freebsd.org (Postfix) with ESMTP id C2E6013C467 for ; Fri, 14 Sep 2007 22:44:37 +0000 (UTC) (envelope-from lists@stringsutils.com) Received: from 35st.simplicato.com (static-71-249-233-130.nycmny.east.verizon.net [71.249.233.130]) by zoraida.natserv.net (Postfix) with ESMTP id 716AE17040; Fri, 14 Sep 2007 18:46:12 -0400 (EDT) References: <1F219879A7E5C565C96109FF@c-2f56e155.1521-1-64736c12.cust.bredbandsbolaget.se> <26F41A5DACB2D2CB43A5829E@c-6254e155.1521-1-64736c12.cust.bredbandsbolaget.se> Message-ID: X-Mailer: http://www.courier-mta.org/cone/ From: Francisco Reyes To: Palle Girgensohn Date: Fri, 14 Sep 2007 18:44:35 -0400 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="US-ASCII" Content-Disposition: inline Content-Transfer-Encoding: 7bit Cc: freebsd-performance@freebsd.org Subject: Re: AMD or Intel? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 14 Sep 2007 22:44:38 -0000 Palle Girgensohn writes: > Sorry, my mistake, more like a percent per day at the moment... If you have 16GB and you got 1% per day.. in less than 100 days you would be at 32GB.. If growth continues you will theoretically be at 60GB+ within a year. > We are planning about 16 GB RAM, actually. Maybe it is overkill? If you can afford it the more the merrier.. given that you likely will have this machine for a good couple of years.. as your data grow what was once "overkill" will become "a good amount" of memory. So I would say do get the 16GB if the budget allows it. > We will probably go for SCSI. HP DL380 with "HP SmartArray", aka ciss. How will you monitor disk failures? Does that controller can be monitored with FreeBSD in some way? From owner-freebsd-performance@FreeBSD.ORG Sat Sep 15 17:50:14 2007 Return-Path: Delivered-To: performance@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id E483816A418; Sat, 15 Sep 2007 17:50:14 +0000 (UTC) (envelope-from kris@FreeBSD.org) Received: from weak.local (hub.freebsd.org [IPv6:2001:4f8:fff6::36]) by mx1.freebsd.org (Postfix) with ESMTP id B140713C442; Sat, 15 Sep 2007 17:50:13 +0000 (UTC) (envelope-from kris@FreeBSD.org) Message-ID: <46EC1B55.4060202@FreeBSD.org> Date: Sat, 15 Sep 2007 19:50:13 +0200 From: Kris Kennaway User-Agent: Thunderbird 2.0.0.6 (Macintosh/20070728) MIME-Version: 1.0 To: java@FreeBSD.org, performance@FreeBSD.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Subject: Massive performance loss from OS::sleep hack X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 15 Sep 2007 17:50:15 -0000 Hi, I have been running the volano java benchmark (http://www.volano.com/benchmarks.html) on an 8-core i386 system, and out of the box jdk15 on FreeBSD performs extremely poorly. The system is more than 90% idle, and profiling shows that the ~800 threads in the benchmark are spending most of their time doing short nanosleep() calls. I traced it to the following FreeBSD-specific hack in the jdk: // XXXBSD: understand meaning and workaround related to yield ... // XXXBSD: done differently in 1.3.1, take a look int os::sleep(Thread* thread, jlong millis, bool interruptible) { assert(thread == Thread::current(), "thread consistency check"); ... if (millis <= 0) { // NOTE: workaround for bug 4338139 if (thread->is_Java_thread()) { ThreadBlockInVM tbivm((JavaThread*) thread); // BSDXXX: Only use pthread_yield here and below if the system thread // scheduler gives time slices to lower priority threads when yielding. #ifdef __FreeBSD__ os_sleep(MinSleepInterval, interruptible); #else pthread_yield(); #endif When I removed this hack (i.e. revert to pthread_yield()) I got an immediate 7-fold performance increase, which brings FreeBSD performance on par with Solaris. What is the reason why this code is necessary? Does FreeBSD's sched_yield() really have different semantics to the other operating systems, or was this a libkse bug that was being worked around? Kris From owner-freebsd-performance@FreeBSD.ORG Sat Sep 15 19:08:32 2007 Return-Path: Delivered-To: performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 98BF416A418; Sat, 15 Sep 2007 19:08:32 +0000 (UTC) (envelope-from kris@FreeBSD.org) Received: from weak.local (hub.freebsd.org [IPv6:2001:4f8:fff6::36]) by mx1.freebsd.org (Postfix) with ESMTP id C92A113C483; Sat, 15 Sep 2007 19:08:29 +0000 (UTC) (envelope-from kris@FreeBSD.org) Message-ID: <46EC2DAA.9070602@FreeBSD.org> Date: Sat, 15 Sep 2007 21:08:26 +0200 From: Kris Kennaway User-Agent: Thunderbird 2.0.0.6 (Macintosh/20070728) MIME-Version: 1.0 To: Daniel Eischen References: <46EC1B55.4060202@FreeBSD.org> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: java@freebsd.org, performance@freebsd.org Subject: Re: Massive performance loss from OS::sleep hack X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 15 Sep 2007 19:08:32 -0000 Daniel Eischen wrote: >> When I removed this hack (i.e. revert to pthread_yield()) I got an >> immediate 7-fold performance increase, which brings FreeBSD >> performance on par with Solaris. >> >> What is the reason why this code is necessary? Does FreeBSD's >> sched_yield() really have different semantics to the other operating >> systems, or was this a libkse bug that was being worked around? > > It's certainly not a libkse bug, at least with scope process threads > libkse does the right thinng. For scope system threads and all libthr > threads, it probably depends on what scheduler you are using since > it's essentially a __sys_sched_yield(). > > On a side note, I think pthread_yield() is deprecated and not in the > latest POSIX spec. sched_yield() is in the spec and is specified > to account for behavior in a threaded environment. Yeah, libthr's pthread_yield just calls sched_yield. Anyway, it seems to me that the decision of what thread to run next is a scheduler decision, and if it is inappropriate for some reason for a scheduler to possibly choose to reschedule the same thread that just yielded, this is something that should be addressed in the scheduler rather than by adding hacks to the application. Kris From owner-freebsd-performance@FreeBSD.ORG Sat Sep 15 18:41:44 2007 Return-Path: Delivered-To: performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 664BF16A417; Sat, 15 Sep 2007 18:41:44 +0000 (UTC) (envelope-from deischen@freebsd.org) Received: from mail.netplex.net (mail.netplex.net [204.213.176.10]) by mx1.freebsd.org (Postfix) with ESMTP id 0784A13C458; Sat, 15 Sep 2007 18:41:43 +0000 (UTC) (envelope-from deischen@freebsd.org) Received: from sea.ntplx.net (sea.ntplx.net [204.213.176.11]) by mail.netplex.net (8.14.1/8.14.1/NETPLEX) with ESMTP id l8FISrp7017490; Sat, 15 Sep 2007 14:28:53 -0400 (EDT) X-Virus-Scanned: by AMaViS and Clam AntiVirus (mail.netplex.net) X-Greylist: Message whitelisted by DRAC access database, not delayed by milter-greylist-3.0 (mail.netplex.net [204.213.176.10]); Sat, 15 Sep 2007 14:28:53 -0400 (EDT) Date: Sat, 15 Sep 2007 14:28:53 -0400 (EDT) From: Daniel Eischen X-X-Sender: eischen@sea.ntplx.net To: Kris Kennaway In-Reply-To: <46EC1B55.4060202@FreeBSD.org> Message-ID: References: <46EC1B55.4060202@FreeBSD.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-Mailman-Approved-At: Sat, 15 Sep 2007 22:37:59 +0000 Cc: java@freebsd.org, performance@freebsd.org Subject: Re: Massive performance loss from OS::sleep hack X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: Daniel Eischen List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 15 Sep 2007 18:41:44 -0000 On Sat, 15 Sep 2007, Kris Kennaway wrote: > Hi, > > I have been running the volano java benchmark > (http://www.volano.com/benchmarks.html) on an 8-core i386 system, and out of > the box jdk15 on FreeBSD performs extremely poorly. The system is more than > 90% idle, and profiling shows that the ~800 threads in the benchmark are > spending most of their time doing short nanosleep() calls. > > > I traced it to the following FreeBSD-specific hack in the jdk: > > // XXXBSD: understand meaning and workaround related to yield > ... > // XXXBSD: done differently in 1.3.1, take a look > int os::sleep(Thread* thread, jlong millis, bool interruptible) { > assert(thread == Thread::current(), "thread consistency check"); > ... > > if (millis <= 0) { > // NOTE: workaround for bug 4338139 > if (thread->is_Java_thread()) { > ThreadBlockInVM tbivm((JavaThread*) thread); > // BSDXXX: Only use pthread_yield here and below if the system thread > // scheduler gives time slices to lower priority threads when yielding. > #ifdef __FreeBSD__ > os_sleep(MinSleepInterval, interruptible); > #else > pthread_yield(); > #endif > > When I removed this hack (i.e. revert to pthread_yield()) I got an immediate > 7-fold performance increase, which brings FreeBSD performance on par with > Solaris. > > What is the reason why this code is necessary? Does FreeBSD's sched_yield() > really have different semantics to the other operating systems, or was this a > libkse bug that was being worked around? It's certainly not a libkse bug, at least with scope process threads libkse does the right thinng. For scope system threads and all libthr threads, it probably depends on what scheduler you are using since it's essentially a __sys_sched_yield(). On a side note, I think pthread_yield() is deprecated and not in the latest POSIX spec. sched_yield() is in the spec and is specified to account for behavior in a threaded environment. -- DE