From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S274965AbTHAWk4 (ORCPT ); Fri, 1 Aug 2003 18:40:56 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S274969AbTHAWk4 (ORCPT ); Fri, 1 Aug 2003 18:40:56 -0400 Received: from main.gmane.org ([80.91.224.249]:64740 "EHLO main.gmane.org") by vger.kernel.org with ESMTP id S274965AbTHAWkx (ORCPT ); Fri, 1 Aug 2003 18:40:53 -0400 X-Injected-Via-Gmane: http://gmane.org/ To: linux-kernel@vger.kernel.org From: Wes Felter Subject: Re: HELP: cpufreq on HT and/or SMP systems Date: Fri, 01 Aug 2003 17:36:21 -0500 Message-ID: References: <200307312353.54735.gallir@uib.es> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT X-Complaints-To: usenet@main.gmane.org User-Agent: Pan/0.14.0 (I'm Being Nibbled to Death by Cats!) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 31 Jul 2003 23:53:54 +0200, Ricardo Galli wrote: > I have some doubts regarding cpufreq for SMP systems (for developing > http://mnm.uib.es/~gallir/cpudyn/). > > If I remember well, I read a while ago that cpufreq didn't work in SMP > systems, but reading the docs and kernel/cpufreq.c, it seems there should be > any problem? > > Is it true? > > OTH, I trying it on a P4 HT, and it works, changes in the frequency of one of > the "cpu's", changes both. That makes sense, since the CPU only has one clock. > What happens in the case of several real cpu's? Does it keep the same > frequency for every cpu? According to a comment in cpufreq.c, it seems that > each cpu might have different frequencies. AFAIK no SMP systems have voltage/frequency scaling (SpeedStep/PowerNow). I've heard that ACPI P-states works on SMP, but if it's not doing voltage/frequency scaling then I don't know what it's doing. Wes Felter - wesley@felter.org - http://felter.org/wesley/