From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761193AbXHTTlp (ORCPT ); Mon, 20 Aug 2007 15:41:45 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1760794AbXHTTlh (ORCPT ); Mon, 20 Aug 2007 15:41:37 -0400 Received: from netops-testserver-4-out.sgi.com ([192.48.171.29]:48402 "EHLO relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1756264AbXHTTlg (ORCPT ); Mon, 20 Aug 2007 15:41:36 -0400 Date: Mon, 20 Aug 2007 12:41:36 -0700 (PDT) From: Christoph Lameter X-X-Sender: clameter@schroedinger.engr.sgi.com To: Mathieu Desnoyers cc: linux-kernel@vger.kernel.org Subject: Re: [PATCH 7/7] Simple Performance Counters: SLUB instrumentation In-Reply-To: <20070818032015.GA28293@Krystal> Message-ID: References: <1185924312685-git-send-email-clameter@sgi.com> <20070817162154.GB30922@Krystal> <20070818032015.GA28293@Krystal> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 17 Aug 2007, Mathieu Desnoyers wrote: > > The cpu will definitely not be idle during these measurements and no > > frequency scaling is active. > > The problem is that if the cpu is idle _before_ the measurements, the > frequency will change differently from one cpu to another. Therefore, > the cycle counters may have large offsets when you start your tests. So, > if get_cycles() is executed on different CPUs (thread being migrated) > between the beginning and the end of the test, the results would be > skewed. TSC measurements as done by this patch are associated with cpus. The skew is irrelevant. Development of this patchset was done on a system whose TSC are never synchronized.