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 Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 3FDD7C433EF for ; Fri, 22 Apr 2022 08:12:02 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1445506AbiDVIOw (ORCPT ); Fri, 22 Apr 2022 04:14:52 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:53968 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233430AbiDVIOu (ORCPT ); Fri, 22 Apr 2022 04:14:50 -0400 Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 6A0CA52B35 for ; Fri, 22 Apr 2022 01:11:58 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1650615118; x=1682151118; h=message-id:subject:from:to:cc:date:in-reply-to: references:mime-version:content-transfer-encoding; bh=cK0QbodZ/3uZEk4psLp/3izywPUoNzicZxBVh2alexk=; b=nRRJugU76KQmSZ0mB8cpff4kJ+dx/HnuFlesyvovbSwNEg3uBEQTNvNm 8ePXKLdW1Drs1Ax0RMUGVqUXduvWrNCJqseTuNepgol6t8BOKThJ2VfFZ yovc0qzCfTegy5JnZHBaYqE+NjQdyuiPfDlauu7GU0PPGtbjCcE15x6Dq nKCYgzQQoUdWfujO8ptt5dyHqxME9IHA9wgZzkLSE7WCK4O17mp37f5GZ VnZsSp2wftDn889SRhooa+b5CnKCQ4+3lQAZv5GllQhOX/R249w/sch77 dUTWYU0r+pMTXvCI/LoQcUlbVOiMVFOzVtjTaW8p69JsaR3P8lh5r5Rul Q==; X-IronPort-AV: E=McAfee;i="6400,9594,10324"; a="264782158" X-IronPort-AV: E=Sophos;i="5.90,281,1643702400"; d="scan'208";a="264782158" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Apr 2022 01:11:58 -0700 X-IronPort-AV: E=Sophos;i="5.90,281,1643702400"; d="scan'208";a="577751144" Received: from jiejingx-mobl1.ccr.corp.intel.com ([10.254.215.31]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Apr 2022 01:11:55 -0700 Message-ID: <8335c2c389c2aee51d91a79bee977f4df6f8fb77.camel@intel.com> Subject: Re: [sched/fair] 2cfb7a1b03: stress-ng.fstat.ops_per_sec -19.8% regression From: "ying.huang@intel.com" To: Mel Gorman Cc: Peter Zijlstra , "Gautham R. Shenoy" , LKML , lkp@lists.01.org, lkp@intel.com, feng.tang@intel.com, zhengjun.xing@linux.intel.com, fengwei.yin@intel.com, aubrey.li@linux.intel.com, yu.c.chen@intel.com, kernel test robot Date: Fri, 22 Apr 2022 16:11:52 +0800 In-Reply-To: References: <20220418141856.GC25584@xsang-OptiPlex-9020> Content-Type: text/plain; charset="UTF-8" User-Agent: Evolution 3.38.3-1 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi, Mel, On Wed, 2022-04-20 at 14:24 +0800, ying.huang@intel.com wrote: > Hi, Mel, > > On Mon, 2022-04-18 at 22:18 +0800, kernel test robot wrote: > > (please be noted we reported "[sched/fair] 2cfb7a1b03: fsmark.files_per_sec > > -26.2% regression" at > > https://lore.kernel.org/all/20220303153108.GC14527@xsang-OptiPlex-9020/ > > when this is still on branch: > > commit: 2cfb7a1b031b0e816af7a6ee0c6ab83b0acdf05a ("sched/fair: Improve consistency of allowed NUMA balance calculations") > > https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git sched/core > > > > now we noticed the similar performance changes as well as some others are > > still existing on mainline, so report this again for information) > > > > > > Greeting, > > > > FYI, we noticed a -19.8% regression of stress-ng.fstat.ops_per_sec due to commit: > > > > > > commit: 2cfb7a1b031b0e816af7a6ee0c6ab83b0acdf05a ("sched/fair: Improve consistency of allowed NUMA balance calculations") > > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master > > > > in testcase: stress-ng > > on test machine: 96 threads 2 sockets Ice Lake with 256G memory > > with following parameters: > > > > nr_threads: 10% > > disk: 1HDD > > testtime: 60s > > fs: f2fs > > class: filesystem > > test: fstat > > cpufreq_governor: performance > > ucode: 0xb000280 > > > > > > In addition to that, the commit also has significant impact on the following tests: > > > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | phoronix-test-suite: phoronix-test-suite.neatbench.CPU.fps 12.5% improvement | > > > test machine | 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with 512G memory | > > > test parameters | cpufreq_governor=performance | > > >                  | option_a=CPU | > > >                  | test=neatbench-1.0.4 | > > >                  | ucode=0x500320a | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | phoronix-test-suite: phoronix-test-suite.neatbench.All.fps 15.2% improvement | > > > test machine | 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with 512G memory | > > > test parameters | cpufreq_governor=performance | > > >                  | option_a=All (CPU + GPU) | > > >                  | test=neatbench-1.0.4 | > > >                  | ucode=0x500320a | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | fsmark: fsmark.files_per_sec -9.9% regression | > > > test machine | 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory | > > > test parameters | cpufreq_governor=performance | > > >                  | disk=1BRD_48G | > > >                  | filesize=4M | > > >                  | fs=f2fs | > > >                  | iterations=1x | > > >                  | nr_threads=64t | > > >                  | sync_method=fsyncBeforeClose | > > >                  | test_size=24G | > > >                  | ucode=0x500320a | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | fsmark: fsmark.files_per_sec -26.2% regression | > > > test machine | 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory | > > > test parameters | cpufreq_governor=performance | > > >                  | disk=1BRD_48G | > > >                  | filesize=4M | > > >                  | fs=ext4 | > > >                  | iterations=1x | > > >                  | nr_threads=64t | > > >                  | sync_method=NoSync | > > >                  | test_size=24G | > > >                  | ucode=0x500320a | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | stress-ng: stress-ng.fstat.ops_per_sec -20.1% regression | > > > test machine | 96 threads 2 sockets Ice Lake with 256G memory | > > > test parameters | class=filesystem | > > >                  | cpufreq_governor=performance | > > >                  | disk=1HDD | > > >                  | fs=xfs | > > >                  | nr_threads=10% | > > >                  | test=fstat | > > >                  | testtime=60s | > > >                  | ucode=0xb000280 | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | fsmark: fsmark.files_per_sec -16.3% regression | > > > test machine | 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory | > > > test parameters | cpufreq_governor=performance | > > >                  | disk=1BRD_48G | > > >                  | filesize=4M | > > >                  | fs=ext4 | > > >                  | iterations=1x | > > >                  | nr_threads=64t | > > >                  | sync_method=fsyncBeforeClose | > > >                  | test_size=24G | > > >                  | ucode=0x500320a | > > +------------------+-------------------------------------------------------------------------------------+ > > > testcase: change | stress-ng: stress-ng.fstat.ops_per_sec -20.2% regression | > > > test machine | 96 threads 2 sockets Ice Lake with 256G memory | > > > test parameters | class=filesystem | > > >                  | cpufreq_governor=performance | > > >                  | disk=1HDD | > > >                  | fs=xfs | > > >                  | nr_threads=10% | > > >                  | test=fstat | > > >                  | testtime=60s | > > >                  | ucode=0xb000280 | > > +------------------+-------------------------------------------------------------------------------------+ > > > > When I worked on the following regression report, > > https://lore.kernel.org/lkml/87tuc7fp9k.fsf@yhuang6-desk2.ccr.corp.intel.com/ > > I found stress-ng throughput will regress if the tasks are distrubuted > more evenly among NUMA nodes. So for this regression, I re-tested with > mpstat per node statistics. The results are as follows, > > mpstat.node.0.user% mpstat.node.1.user% mpstat.node.0.sys% mpstat.node.1.sys% > 889c5d60fb 3.04 2.65 30.0 25.8 > 2cfb7a1b03 2.39 2.28 31.2 28.9 > > It can be found that the task are balanced better with the commit > 2cfb7a1b03. So I think the regression isn't a real problem. I think stress-ng isn't a good workload to evaluate the effect of this patch. Can you teach me which workloads are appropriate? Best Regards, Huang, Ying