From owner-cvs-all Wed Jan 29 17:27:13 2003 Delivered-To: cvs-all@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 99B0837B401; Wed, 29 Jan 2003 17:27:11 -0800 (PST) Received: from canning.wemm.org (canning.wemm.org [192.203.228.65]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4FDF743F43; Wed, 29 Jan 2003 17:27:11 -0800 (PST) (envelope-from peter@wemm.org) Received: from wemm.org (localhost [127.0.0.1]) by canning.wemm.org (Postfix) with ESMTP id 073622A89E; Wed, 29 Jan 2003 17:27:06 -0800 (PST) (envelope-from peter@wemm.org) X-Mailer: exmh version 2.5 07/13/2001 with nmh-1.0.4 To: Lars Eggert Cc: Dmitry Morozovsky , cvs-all@FreeBSD.ORG, Nate Lawson Subject: Re: cvs commit: src/sys/dev/acpica acpi_cpu.c In-Reply-To: <3E2545E9.3080803@isi.edu> Date: Wed, 29 Jan 2003 17:27:06 -0800 From: Peter Wemm Message-Id: <20030130012706.073622A89E@canning.wemm.org> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Lars Eggert wrote: > On Tue, 14 Jan 2003, Peter Wemm wrote: > > > > FWIW, we see a (measured) difference of about 50W on 2.2GHz P4's > > simply by turning machdep.cpu_idle_hlt on and off. I expect the > > clock throttling would make similar differences. For 1U > > rack-mount systems (especially in California) this is a Big Deal. > > Since the AC in my office has a hard time with four new Dell SMPs, would > setting machdep.cpu_idle_hlt=1 translate to making them run cooler? They > have a pretty bursty workload, so many times they're idle, and sometimes > fully loaded - does the hlt setting affect performance? This is a bit old, but it can cause a slight slowdown, because it may take a little longer for a cpu to notice a job sitting on the run queue that it could have been running. Once it does a halt, it will stop checking until the next interrupt (sometimes a clock, sometimes disk/net/etc). This usually works out to be a slight loss, but has been attributed to a slight improvement in a couple of cases by causing less cache coherency bus thrashing etc. This however is highly dependent on the cpu bus architecture. It will cool things down somewhat, but *only when the machine is idle*. If it is under full load for an hour or so, your AC will still need to be able to deal with that. But it will be a big win for less than maxed out machines or bursty machines. Cheers, -Peter -- Peter Wemm - peter@wemm.org; peter@FreeBSD.org; peter@yahoo-inc.com "All of this is for nothing if we don't go to the stars" - JMS/B5 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message