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,URIBL_BLOCKED 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 183E6C433F5 for ; Mon, 10 Sep 2018 10:40:23 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id BD1F22086E for ; Mon, 10 Sep 2018 10:40:22 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org BD1F22086E 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 S1727872AbeIJPdq (ORCPT ); Mon, 10 Sep 2018 11:33:46 -0400 Received: from mga17.intel.com ([192.55.52.151]:31891 "EHLO mga17.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727353AbeIJPdp (ORCPT ); Mon, 10 Sep 2018 11:33:45 -0400 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga005.jf.intel.com ([10.7.209.41]) by fmsmga107.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 10 Sep 2018 03:40:20 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.53,355,1531810800"; d="scan'208";a="255916451" Received: from linux.intel.com ([10.54.29.200]) by orsmga005.jf.intel.com with ESMTP; 10 Sep 2018 03:40:20 -0700 Received: from [10.125.252.9] (abudanko-mobl.ccr.corp.intel.com [10.125.252.9]) by linux.intel.com (Postfix) with ESMTP id 606B25803DA; Mon, 10 Sep 2018 03:40:18 -0700 (PDT) Subject: Re: [PATCH v8 0/3]: perf: reduce data loss when profiling highly parallel CPU bound workloads To: Ingo Molnar Cc: Peter Zijlstra , Arnaldo Carvalho de Melo , Alexander Shishkin , Jiri Olsa , Namhyung Kim , Andi Kleen , linux-kernel References: <20180910091841.GA4664@gmail.com> From: Alexey Budankov Organization: Intel Corp. Message-ID: <2c5d4b01-0eb8-f97e-6a70-44be7961d7f8@linux.intel.com> Date: Mon, 10 Sep 2018 13:40:17 +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: <20180910091841.GA4664@gmail.com> 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 Ingo, On 10.09.2018 12:18, Ingo Molnar wrote: > > * 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. > > Yay! I saw this frequently on a 120-CPU box (hw is broken now). > >> 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. >> >> Applying asynchronous trace streaming thru Posix AIO API >> (http://man7.org/linux/man-pages/man7/aio.7.html) >> lowers data loss metrics value providing 2x improvement - >> lowering 98% loss to almost 0%. > > Hm, instead of AIO why don't we use explicit threads instead? I think Posix AIO will fall back > to threads anyway when there's no kernel AIO support (which there probably isn't for perf > events). Explicit threading is surely an option but having more threads in the tool that stream performance data is a considerable design complication. Luckily, glibc AIO implementation is already based on pthreads, but having a writing thread for every distinct fd only. > > Per-CPU threading the record session would have so many other advantages as well (scalability, > etc.).> > Jiri did per-CPU recording patches a couple of months ago, not sure how usable they are at the > moment? Tool threads may contend, and actually do, with application threads, under heavy load when all CPU cores are utilized, and this may alter performance profile. Thanks, Alexey > > Thanks, > > Ingo >