From owner-freebsd-questions@FreeBSD.ORG Wed Sep 28 19:23:52 2005 Return-Path: X-Original-To: freebsd-questions@freebsd.org Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8531F16A41F for ; Wed, 28 Sep 2005 19:23:52 +0000 (GMT) (envelope-from tijl@ulyssis.org) Received: from nibbel.kulnet.kuleuven.ac.be (nibbel.kulnet.kuleuven.ac.be [134.58.240.41]) by mx1.FreeBSD.org (Postfix) with ESMTP id DF75C43D48 for ; Wed, 28 Sep 2005 19:23:51 +0000 (GMT) (envelope-from tijl@ulyssis.org) Received: from localhost (localhost [127.0.0.1]) by nibbel.kulnet.kuleuven.ac.be (Postfix) with ESMTP id 905364C17A; Wed, 28 Sep 2005 21:23:50 +0200 (CEST) Received: from smtp02.kuleuven.be (lepidus.kulnet.kuleuven.ac.be [134.58.240.72]) by nibbel.kulnet.kuleuven.ac.be (Postfix) with ESMTP id DE8B94C14E; Wed, 28 Sep 2005 21:23:49 +0200 (CEST) Received: from kalimero.kotnet.org (kalimero.kotnet.org [10.4.42.180]) by smtp02.kuleuven.be (Postfix) with ESMTP id 964A92CAA3B; Wed, 28 Sep 2005 21:23:48 +0200 (CEST) Received: from kalimero.kotnet.org (kalimero.kotnet.org [127.0.0.1]) by kalimero.kotnet.org (8.13.4/8.13.3) with ESMTP id j8SJNjTp014528; Wed, 28 Sep 2005 21:23:46 +0200 (CEST) (envelope-from tijl@ulyssis.org) From: Tijl Coosemans To: freebsd-questions@freebsd.org Date: Wed, 28 Sep 2005 21:23:44 +0200 User-Agent: KMail/1.8.2 References: <200509281449.40834@harrymail> In-Reply-To: <200509281449.40834@harrymail> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200509282123.44837.tijl@ulyssis.org> X-Virus-Scanned: by KULeuven Antivirus Cluster Cc: Emanuel Strobl Subject: Re: acpi_thorttle and cpufreq/ichss questions X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 28 Sep 2005 19:23:52 -0000 On Wednesday 28 September 2005 14:49, Emanuel Strobl wrote: > my laptop (PIII LV) runs fine with cpufreq and ichss: > > dev.ichss.0.%desc: SpeedStep ICH > dev.ichss.0.%driver: ichss > dev.ichss.0.%parent: cpu0 > dev.ichss.0.freq_settings: 794/-1 497/-1 > > If I understand the output correctly i have two SpeedSteps, the full > 800MHz (with higher voltage) and 500MHz. Correct. > But now I also have acpi_throttle (if I don't disable it > in /boot/loader.conf): > > dev.acpi_throttle.0.%desc: ACPI CPU Throttling > dev.acpi_throttle.0.%driver: acpi_throttle > dev.acpi_throttle.0.%parent: cpu0 > dev.acpi_throttle.0.freq_settings: 10000/-1 8750/-1 7500/-1 6250/-1 > 5000/-1 3750/-1 2500/-1 1250/-1 > > What do these values tell me? I think that are divisors, but how can > I see which one is used? These are percentage values. 10000 is 100.00%, 8750 is 87.50% etc. Currently there doesn't seem to be a way to know the current setting of a specific cpufreq driver. You could however derive it from dev.cpu.0.freq. > Can I constrain the divisors, lets say I want to have only 10000, > 7500 and 5000? (I know I can limit the lowest frequency cpufreq may > use with "debug.cpufreq.lowest=310" but then I don't know if it > throttles the 800MHz mode (thus using high core voltage) or if it > throttles the power saving 500MHz mode) Cpufreq will not throttle 800MHz below 500MHz. For example, 800*50% should be missing from dev.cpu.0.freq_levels. > And what does this tell me: > > hw.acpi.cpu.cx_supported: C1/0 C2/10 C3/85 > hw.acpi.cpu.cx_lowest: C1 > hw.acpi.cpu.cx_usage: 100.00% 0.00% 0.00% This is something different entirely. Cx states are CPU power states used when the CPU is idle. The higher, the less power it consumes, but the longer it takes to return from that state. I don't know what the exact differences are, but C1 is probably a simple halt instruction, whereas C2 and C3 probably also disable things like caches. I suppose hw.acpi.cpu.cx_lowest allows you to set the lowest Cx state the OS (scheduler) should use, but I haven't experimented with this myself yet because of errors in my laptop's bios...