Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263226AbUCYQYe (ORCPT ); Thu, 25 Mar 2004 11:24:34 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263234AbUCYQYe (ORCPT ); Thu, 25 Mar 2004 11:24:34 -0500 Received: from mtvcafw.SGI.COM ([192.48.171.6]:31138 "EHLO omx3.sgi.com") by vger.kernel.org with ESMTP id S263226AbUCYQYd (ORCPT ); Thu, 25 Mar 2004 11:24:33 -0500 Message-ID: <004101c41285$4e820270$6800a8c0@comcast.net> From: "John Hawkes" To: "Nakajima, Jun" , "Andi Kleen" , "Ingo Molnar" Cc: , , , , , , , , References: <7F740D512C7C1046AB53446D3720017301119907@scsmsx402.sc.intel.com> Subject: Re: [Lse-tech] [patch] sched-domain cleanups, sched-2.6.5-rc2-mm2-A3 Date: Thu, 25 Mar 2004 08:19:28 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6600 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1076 Lines: 22 From: "Nakajima, Jun" > We have found some performance regressions (e.g. SPECjbb) with the > scheduler on a large IA-64 NUMA machine, and we are debugging it. On SMP > machines, we haven't seen performance regressions. I've run a flavor of AIM7 and kernbench on a variety of ia64-NUMA CPU counts, ranging up to 128p (and my 2.6.4 + Piggin-scheduler kernel hangs during boot at >=192p, vs. 2.6.4 + vanilla-scheduler booting and running at 512p), and the Piggin scheduler is 10-15% slower at 64p, and even worse at 128p. I did, however, produce superior performance (superior to the vanilla scheduler) with AIM7 by increasing the sched_domain->busy_factor by 4x. I tried a few other things, but nothing improved performance on these big systems better than a simple increase of busy_factor. John Hawkes - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/