From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-0.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E7B7DC433F4 for ; Tue, 28 Aug 2018 11:58:33 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 9329D2087A for ; Tue, 28 Aug 2018 11:58:33 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 9329D2087A Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727456AbeH1Ptu (ORCPT ); Tue, 28 Aug 2018 11:49:50 -0400 Received: from mga01.intel.com ([192.55.52.88]:5318 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727085AbeH1Ptu (ORCPT ); Tue, 28 Aug 2018 11:49:50 -0400 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 28 Aug 2018 04:58:30 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.53,299,1531810800"; d="scan'208";a="80840302" Received: from linux.intel.com ([10.54.29.200]) by fmsmga002.fm.intel.com with ESMTP; 28 Aug 2018 04:58:30 -0700 Received: from [10.125.252.157] (abudanko-mobl.ccr.corp.intel.com [10.125.252.157]) by linux.intel.com (Postfix) with ESMTP id 4F8A35802AD; Tue, 28 Aug 2018 04:58:29 -0700 (PDT) Subject: Re: [PATCH v3 0/2]: perf: reduce data loss when profiling highly parallel CPU bound workloads To: Andi Kleen Cc: Arnaldo Carvalho de Melo , Alexander Shishkin , Namhyung Kim , Andi Kleen , linux-kernel References: <20180828085904.GH23727@krava> From: Alexey Budankov Organization: Intel Corp. Message-ID: Date: Tue, 28 Aug 2018 14:58:28 +0300 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <20180828085904.GH23727@krava> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Andi, On 28.08.2018 11:59, Jiri Olsa wrote: > On Mon, Aug 27, 2018 at 08:03:21PM +0300, Alexey Budankov wrote: >> >> Currently in record mode the tool implements trace writing serially. >> The algorithm loops over mapped per-cpu data buffers and stores ready >> data chunks into a trace file using write() system call. >> >> At some circumstances the kernel may lack free space in a buffer >> because the other buffer's half is not yet written to disk due to >> some other buffer's data writing by the tool at the moment. >> >> Thus serial trace writing implementation may cause the kernel >> to loose profiling data and that is what observed when profiling >> highly parallel CPU bound workloads on machines with big number >> of cores. >> >> Experiment with profiling matrix multiplication code executing 128 >> threads on Intel Xeon Phi (KNM) with 272 cores, like below, >> demonstrates data loss metrics value of 98%: >> >> /usr/bin/time perf record -o /tmp/perf-ser.data -a -N -B -T -R -g \ >> --call-graph dwarf,1024 --user-regs=IP,SP,BP \ >> --switch-events -e cycles,instructions,ref-cycles,software/period=1,name=cs,config=0x3/Duk -- \ >> matrix.gcc >> >> Data loss metrics is the ratio lost_time/elapsed_time where >> lost_time is the sum of time intervals containing PERF_RECORD_LOST >> records and elapsed_time is the elapsed application run time >> under profiling. > > I like the idea and I think it's good direction to go, but could > you please share some from perf stat or whatever you used to meassure > the new performance? Is it ok to share VTune GUI screenshots I sent you the last time to demonstrate the advantage of AIO trace streaming? Thanks, Alexey > > thanks, > jirka >