From owner-freebsd-stable@FreeBSD.ORG Tue Apr 20 07:56:40 2004 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6B0D516A4CE for ; Tue, 20 Apr 2004 07:56:40 -0700 (PDT) Received: from cray.e-card.bg (mjak.e-card.bg [212.91.167.5]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1FA9D43D4C for ; Tue, 20 Apr 2004 07:56:36 -0700 (PDT) (envelope-from altares@cray.e-card.bg) Received: from cray.e-card.bg (localhost [127.0.0.1]) by cray.e-card.bg (8.12.9/8.12.9) with ESMTP id i3KEvr5k069925; Tue, 20 Apr 2004 17:57:53 +0300 (EEST) (envelope-from altares@cray.e-card.bg) Received: (from altares@localhost) by cray.e-card.bg (8.12.9/8.12.9/Submit) id i3KEvoGC069924; Tue, 20 Apr 2004 17:57:50 +0300 (EEST) Date: Tue, 20 Apr 2004 17:57:50 +0300 From: Rumen Telbizov To: Pete French Message-ID: <20040420145750.GA32493@e-card.bg> References: <20040420073837.GS32493@e-card.bg> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.1i cc: stable@freebsd.org Subject: Re: HTT and SMP question X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 20 Apr 2004 14:56:40 -0000 On Tue, Apr 20, 2004 at 02:48:26PM +0100, Pete French wrote: > > Would you please point out some cases when HTT with machdep.hlt_logical_cpus=0 > > causes worse performance than machdep.hlt_logical_cpus=1? > > Doing a large 'make' does it for me - small difference, but measurable. > I get the best poerformance by compiling out SMP alltogether though. > As my machine spends most of its time doing 'make' then for me it's worth Ok > removing it entirely (aside from the fact that my machine isnt stable > under HTT) I have a single Xeon 2.66 on a dual processor board and I did recompile the kernel with SMP support in first place! HTT is enabled by default! With these 2 things in mind - the box is rock solid for over 2 weeks with average load. I am not sure about _how much_ faster it is with HTT (and machdep.hlt_logical_cpus=0) but at least it IS stable! The OS is 4.9-STABLE (cvsuped 1 week ago). Are you sure that exactly the HTT is causing you the instabillity problems ? Regards Rumen Telbizov