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=-4.0 required=3.0 tests=BAYES_00,MAILING_LIST_MULTI, SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED autolearn=no 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 A3FC9C433E6 for ; Thu, 18 Mar 2021 04:33:27 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 61F0F64F10 for ; Thu, 18 Mar 2021 04:33:27 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229664AbhCREcz (ORCPT ); Thu, 18 Mar 2021 00:32:55 -0400 Received: from mail-lf1-f51.google.com ([209.85.167.51]:42510 "EHLO mail-lf1-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229584AbhCREcn (ORCPT ); Thu, 18 Mar 2021 00:32:43 -0400 Received: by mail-lf1-f51.google.com with SMTP id o10so2398566lfb.9 for ; Wed, 17 Mar 2021 21:32:42 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=boH6m1jSA/WN9PxrEykfEhqNhVP+e9nXbufeG1rDlxg=; b=ZHInkikUjjd5haEJJ8Ai7CVPCK136bGOmoTK7JV/kL3Aon7nldbvS+RyJix5L7Kqr7 6VIQVrpZdVcUEPUV08t07ZkClc+2iamBX580DI6ZDcGXFA8OSheR1v3UziPg9HdZ60Bt qzHem+Rtmdy43hqgZonunlwUoZYN3B2n1ivP1PBWX8mKjaMQnuRIgCfpNsvXkyGj/wig M686SygI/46WpgpdrpXrrAM+cSZifZs0Duoy1bf3/phe/gXiV9EqAS3rKj0eWlNDlRKS 4CRXCKf/6awY9cE/pOEIjRzCb/5CwZdBxnJML5S6mDK4vAg3KWojb6fQykv76rxYUvFc PNtg== X-Gm-Message-State: AOAM533L3HGC02Dr05amL98ca0JvOldhNglloJm+6dLavJuV74+YSJgl 1RghhkU4TPtiiFAH2GXU811VYfQDA5C/vVucSp3O2DdH X-Google-Smtp-Source: ABdhPJyHa/urGDQlvOW8jDAzRusi83V0+t0Zy/jytarlCErtdKNN9ykBZ5WKEabkazOaio06hNDfwdlCyoxLXMRNAxQ= X-Received: by 2002:ac2:5093:: with SMTP id f19mr4389827lfm.494.1616041961766; Wed, 17 Mar 2021 21:32:41 -0700 (PDT) MIME-Version: 1.0 References: <20210316211837.910506-1-songliubraving@fb.com> <7D48A756-C253-48DE-B536-826314778404@fb.com> In-Reply-To: <7D48A756-C253-48DE-B536-826314778404@fb.com> From: Namhyung Kim Date: Thu, 18 Mar 2021 13:32:30 +0900 Message-ID: Subject: Re: [PATCH v2 0/3] perf-stat: share hardware PMCs with BPF To: Song Liu Cc: Arnaldo Carvalho de Melo , linux-kernel , Kernel Team , Arnaldo Carvalho de Melo , Jiri Olsa Content-Type: text/plain; charset="UTF-8" Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Mar 18, 2021 at 12:52 PM Song Liu wrote: > > > > > On Mar 17, 2021, at 6:11 AM, Arnaldo Carvalho de Melo wrote: > > > > Em Wed, Mar 17, 2021 at 02:29:28PM +0900, Namhyung Kim escreveu: > >> Hi Song, > >> > >> On Wed, Mar 17, 2021 at 6:18 AM Song Liu wrote: > >>> > >>> perf uses performance monitoring counters (PMCs) to monitor system > >>> performance. The PMCs are limited hardware resources. For example, > >>> Intel CPUs have 3x fixed PMCs and 4x programmable PMCs per cpu. > >>> > >>> Modern data center systems use these PMCs in many different ways: > >>> system level monitoring, (maybe nested) container level monitoring, per > >>> process monitoring, profiling (in sample mode), etc. In some cases, > >>> there are more active perf_events than available hardware PMCs. To allow > >>> all perf_events to have a chance to run, it is necessary to do expensive > >>> time multiplexing of events. > >>> > >>> On the other hand, many monitoring tools count the common metrics (cycles, > >>> instructions). It is a waste to have multiple tools create multiple > >>> perf_events of "cycles" and occupy multiple PMCs. > >> > >> Right, it'd be really helpful when the PMCs are frequently or mostly shared. > >> But it'd also increase the overhead for uncontended cases as BPF programs > >> need to run on every context switch. Depending on the workload, it may > >> cause a non-negligible performance impact. So users should be aware of it. > > > > Would be interesting to, humm, measure both cases to have a firm number > > of the impact, how many instructions are added when sharing using > > --bpf-counters? > > > > I.e. compare the "expensive time multiplexing of events" with its > > avoidance by using --bpf-counters. > > > > Song, have you perfmormed such measurements? > > I have got some measurements with perf-bench-sched-messaging: > > The system: x86_64 with 23 cores (46 HT) > > The perf-stat command: > perf stat -e cycles,cycles,instructions,instructions,ref-cycles,ref-cycles > > The benchmark command and output: > ./perf bench sched messaging -g 40 -l 50000 -t > # Running 'sched/messaging' benchmark: > # 20 sender and receiver threads per group > # 40 groups == 1600 threads run > Total time: 10X.XXX [sec] > > > I use the "Total time" as measurement, so smaller number is better. > > For each condition, I run the command 5 times, and took the median of > "Total time". > > Baseline (no perf-stat) 104.873 [sec] > # global > perf stat -a 107.887 [sec] > perf stat -a --bpf-counters 106.071 [sec] > # per task > perf stat 106.314 [sec] > perf stat --bpf-counters 105.965 [sec] > # per cpu > perf stat -C 1,3,5 107.063 [sec] > perf stat -C 1,3,5 --bpf-counters 106.406 [sec] > > From the data, --bpf-counters is slightly better than the regular event > for all targets. I noticed that the results are not very stable. There > are a couple 108.xx runs in some of the conditions (w/ and w/o > --bpf-counters). Hmm.. so this result is when multiplexing happened, right? I wondered how/why the regular perf stat is slower.. Thanks, Namhyung > > > I also measured the average runtime of the BPF programs, with > > sysctl kernel.bpf_stats_enabled=1 > > For each event, if we have one leader and two followers, the total run > time is about 340ns. IOW, 340ns for two perf-stat reading instructions, > 340ns for two perf-stat reading cycles, etc. > > Thanks, > Song