From owner-freebsd-mobile@FreeBSD.ORG Sun Feb 19 17:31:41 2006 Return-Path: X-Original-To: mobile@freebsd.org Delivered-To: freebsd-mobile@FreeBSD.ORG Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E522D16A420 for ; Sun, 19 Feb 2006 17:31:40 +0000 (GMT) (envelope-from oberman@es.net) Received: from postal4.es.net (postal4.es.net [198.124.252.66]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7255443D49 for ; Sun, 19 Feb 2006 17:31:39 +0000 (GMT) (envelope-from oberman@es.net) Received: from ptavv.es.net ([198.128.4.29]) by postal4.es.net (Postal Node 4) with ESMTP (SSL) id IBA74465; Sun, 19 Feb 2006 09:31:38 -0800 Received: from ptavv.es.net (localhost [127.0.0.1]) by ptavv.es.net (Tachyon Server) with ESMTP id 73F1145041; Sun, 19 Feb 2006 09:31:35 -0800 (PST) To: "M. Warner Losh" In-reply-to: Your message of "Sat, 18 Feb 2006 23:22:30 MST." <20060218.232230.37239553.imp@bsdimp.com> Date: Sun, 19 Feb 2006 09:31:35 -0800 From: "Kevin Oberman" Message-Id: <20060219173135.73F1145041@ptavv.es.net> Cc: anderson@centtech.com, mobile@freebsd.org Subject: Re: i8000fan for freebsd and Sony PCG-z1wa X-BeenThere: freebsd-mobile@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Mobile computing with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 19 Feb 2006 17:31:41 -0000 > Date: Sat, 18 Feb 2006 23:22:30 -0700 (MST) > From: "M. Warner Losh" > > In message: <20060219051441.148E345041@ptavv.es.net> > "Kevin Oberman" writes: > : > Date: Sat, 18 Feb 2006 21:59:35 -0700 (MST) > : > From: "M. Warner Losh" > : > Sender: owner-freebsd-mobile@freebsd.org > : > > : > In message: <43F7F2FD.9070509@centtech.com> > : > Eric Anderson writes: > : > : M. Warner Losh wrote: > : > : > Recently, I've had to diagnose problems with my dell i8200 which sadly > : > : > runs windows. It was running slower than the dickens. Turns out it > : > : > was heat related and the bios was stepping the speed down and never > : > : > back up. I've not been able to fix the overheading problem (would > : > : > love to know how, btw). > : > : > > : > : > In the process of all of this, I found a damn useful program that > : > : > monitored the temperature, fan speed CPU load and CPU speed, producing > : > : > a nice graph over time. > : > : > > : > : > I was wondering if something similar existed for FreeBSD. I'd like a > : > : > nice little program that I can use to graph the temperature, CPU speed > : > : > and cpu load, with and without powerd running. Can anybody help me out? > : > : > > : > : > : > : Are you looking for a long term statistical tool, or a real-time > : > : graphical view? I was thinking perl+rrdtool for a long term background > : > : tool that would create png's would be pretty easy to whip up.. > : > > : > I was hoping to get a nice real-time graph... > : > : I have not looked at details, but it looks like whipping up a plugin for > : gkrellm for this would be pretty straight forward. It has all of the > : basic tools to do what you want. It already monitors environment from > : ACPI and can monitor temperatures, voltages and fans using mbmon if the > : information is not available in ACPI. > > How do I get the current speed of the CPU? The rest I think I can do > already... Is it just the dev.cpu.0.freq? Is there some way I can > test this value easily to make sure that the OS' idea of the speed and > the CPU's idea of the speed are the same? > > I've seen a lot of interesting graphs in the windows arena that seem > like they would help battery life a lot w/o compromising performance > all that much that I'm thinking would be fun to code up... There is a plugin for gkrallm that will provide the current clock speed, gkx86info2. It does NOT report throttling...only actual clock speed changes. dev.cpu.0.freq reports the effective speed which is a combination of clock speed and throttling. To get a REALLY slow system, it's far more likely that throttling is kicking in as its effect on performance is dar greater than the reduced clock. Pre-Pentium-M, CPUs only supported 2 clock speeds, normally full speed and 2/3 speed. But throttling could reduce the effective speed to 12.5% of the clock speed which can be a real pool of molasses. dev.cpu.0.freq will reflect the effect of thermal management, so will let you know if the system is slowing because it is too hot (or thinks it is). hw.acpi.thermal will tell you where thermal management takes over (_PSV). -- R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634