Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267515AbUIGDR0 (ORCPT ); Mon, 6 Sep 2004 23:17:26 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267519AbUIGDR0 (ORCPT ); Mon, 6 Sep 2004 23:17:26 -0400 Received: from relay.pair.com ([209.68.1.20]:39186 "HELO relay.pair.com") by vger.kernel.org with SMTP id S267515AbUIGDRY (ORCPT ); Mon, 6 Sep 2004 23:17:24 -0400 X-pair-Authenticated: 66.188.111.210 Message-ID: <413D2842.8030401@cybsft.com> Date: Mon, 06 Sep 2004 22:17:22 -0500 From: "K.R. Foley" User-Agent: Mozilla Thunderbird 0.7.3 (X11/20040803) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Lee Revell CC: Ingo Molnar , Florian Schmidt , linux-kernel , felipe_alfaro@linuxmail.org Subject: Re: [patch] voluntary-preempt-2.6.9-rc1-bk12-R5 References: <20040903120957.00665413@mango.fruits.de> <20040904195141.GA6208@elte.hu> <20040905140249.GA23502@elte.hu> <1094408203.4445.5.camel@krustophenia.net> <20040905191227.GA29797@elte.hu> <1094418192.4445.58.camel@krustophenia.net> <20040906063040.GA11541@elte.hu> <1094456653.29921.45.camel@krustophenia.net> In-Reply-To: <1094456653.29921.45.camel@krustophenia.net> X-Enigmail-Version: 0.85.0.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1421 Lines: 35 Lee Revell wrote: > On Mon, 2004-09-06 at 02:30, Ingo Molnar wrote: > >>* Lee Revell wrote: >> >> >>>http://krustophenia.net/testresults.php?dataset=2.6.9-rc1-R0#/var/www/2.6.9-rc1-R0/foo.hist >>> >>>I find the two smaller spikes to either side of the central spike >>>really odd. These showed up in my jackd tests too, I had attributed >>>them to some measurement artifact, but they seem real. Maybe a >>>rounding bug, or some kind of weird cache effect? >> >>interesting - the histograms are pretty symmetric around the center. >>E.g. the exponential foo.hist2 diagram is way too symmetric around 50 >>usecs! What precisely is being measured? >> > > > Here's the program. It does mlockall(), acquires realtime scheduling, > then sets up a 2048 Hz stream of interupts from the RTC and measures the > delay. It's quite possible there's a bug, the amlat program did not > seem to work, something must have changed with the RTC from 2.4 to 2.6. > Actually the amlat program works fine for applying real-time scheduling pressure. I believe it just doesn't do any real latency measuring without the hooks provided by Andrew's rtc-debug patch. kr - 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/