From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756254AbZHFQ7g (ORCPT ); Thu, 6 Aug 2009 12:59:36 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753802AbZHFQ7f (ORCPT ); Thu, 6 Aug 2009 12:59:35 -0400 Received: from mail3-relais-sop.national.inria.fr ([192.134.164.104]:64580 "EHLO mail3-relais-sop.national.inria.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752833AbZHFQ7e (ORCPT ); Thu, 6 Aug 2009 12:59:34 -0400 X-IronPort-AV: E=Sophos;i="4.43,334,1246831200"; d="scan'208";a="32155241" Message-ID: <4A7B0C02.7060704@inria.fr> Date: Thu, 06 Aug 2009 18:59:46 +0200 From: Brice Goglin User-Agent: Mozilla-Thunderbird 2.0.0.19 (X11/20090103) MIME-Version: 1.0 To: Ingo Molnar CC: Peter Zijlstra , paulus@samba.org, LKML Subject: Re: [perf] howto switch from pfmon References: <4A3FEF75.2020804@inria.fr> <20090623131450.GA31519@elte.hu> <20090623134749.GA6897@elte.hu> <4A40DFF5.7010207@inria.fr> <20090623143601.GA13415@elte.hu> <4A40F31F.4030609@inria.fr> <20090629192913.GA29295@elte.hu> In-Reply-To: <20090629192913.GA29295@elte.hu> X-Enigmail-Version: 0.95.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Ingo Molnar wrote: > Update: based on your feedback the latest perfcounters tree includes > the following new perf record features: > > -s, --stat per thread counts > -n, --no-samples don't sample > > --stat instructs the kernel to gather precise per task/thread stats > and emits those counts to the data file. Via --no-samples one can do > non-profiling runs - i.e. only statistics collection. > > The 'perf stat' pretty printing side is not fully implemented yet - > right now you can only see these stats if you look for > PERF_EVENT_READ counts in the raw event log: > > perf report -D | grep PERF_EVENT_READ > Ok I am getting remote/local accesses for my threads now. But I am not sure which line corresponds to which event. $ /mnt/scratch/bgoglin/cpunode/linux-2.6.31/tools/perf/perf record -f -s -e r1000001e0 -e r1000002e0 -e r1000004e0 -e r1000008e0 ./stream [...] $ /mnt/scratch/bgoglin/cpunode/linux-2.6.31/tools/perf/perf report -D | grep PERF_EVENT_READ 0x7cd0 [0x30]: PERF_EVENT_READ: 4651 4651 210827 0x7388 [0x30]: PERF_EVENT_READ: 4651 4651 241742 0x8cf0 [0x30]: PERF_EVENT_READ: 4651 4651 315938 0x9da0 [0x30]: PERF_EVENT_READ: 4651 4651 9461794 0x7208 [0x30]: PERF_EVENT_READ: 4651 4652 24954 0x8c90 [0x30]: PERF_EVENT_READ: 4651 4652 408056 0x7ca0 [0x30]: PERF_EVENT_READ: 4651 4652 8962423 0x9ce0 [0x30]: PERF_EVENT_READ: 4651 4652 9117 0x7df0 [0x30]: PERF_EVENT_READ: 4651 4653 21645 0x9d70 [0x30]: PERF_EVENT_READ: 4651 4653 23606 0x7358 [0x30]: PERF_EVENT_READ: 4651 4653 29266 0x8e70 [0x30]: PERF_EVENT_READ: 4651 4653 9339173 [...] I can easily sort them by thread id, but I don't know how to match my 4 events with each group of 4 line. Maybe perf report earned some better way to show per-thread statistics in the meantime? Brice