From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758274AbZENRXh (ORCPT ); Thu, 14 May 2009 13:23:37 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752510AbZENRX2 (ORCPT ); Thu, 14 May 2009 13:23:28 -0400 Received: from relay2.sgi.com ([192.48.179.30]:48250 "EHLO relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752468AbZENRX2 (ORCPT ); Thu, 14 May 2009 13:23:28 -0400 Date: Thu, 14 May 2009 12:23:28 -0500 From: Dimitri Sivanich To: Thomas Gleixner Cc: LKML , Andrew Morton , Peter Zijlstra , Ingo Molnar Subject: Re: [patch 0/2] sched, timers: simplify calc_load and avoid cpu iteration loop Message-ID: <20090514172328.GA25691@sgi.com> References: <20090514105915.973318588@linutronix.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090514105915.973318588@linutronix.de> User-Agent: Mutt/1.5.17 (2007-11-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org I ran this patchset on a 64p ia64 machine and latency-wise it looks good. I can't really speak for the loadavg calculation. On Thu, May 14, 2009 at 11:21:14AM -0000, Thomas Gleixner wrote: > Dimitri Sivanich pointed out that calc_load iterates over all online > CPUs under xtime lock which can cause long latencies on large SMP > systems. > > The following patch series removes the iteration loop and lets > scheduler_tick() on each CPU update the number of active tasks. The > calc_load() function just uses this variable and updates avenrun from > do_timer(). > > Dimitri, can you please test run this on one of your large machines ? > > Thanks, > > tglx >