From mboxrd@z Thu Jan  1 00:00:00 1970
Content-Type: multipart/mixed; boundary="===============8337944977676653964=="
MIME-Version: 1.0
From: Huang, Ying <ying.huang@intel.com>
To: lkp@lists.01.org
Subject: Re: [net] ceb5d58b21: -69.2% fsmark.files_per_sec
Date: Mon, 28 Dec 2015 08:58:17 +0800
Message-ID: <8760zj8k46.fsf@yhuang-dev.intel.com>
In-Reply-To: <CANn89iJxF_qJwk0AmFVd57oexPzeB596KRL0r1HehP23JQM-dw@mail.gmail.com>
List-Id: <oe-lkp.lists.linux.dev>

--===============8337944977676653964==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

Eric Dumazet <edumazet@google.com> writes:

> On Wed, Dec 23, 2015 at 9:49 PM, kernel test robot
> <ying.huang@linux.intel.com> wrote:
>> FYI, we noticed the below changes on
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
>> commit ceb5d58b217098a657f3850b7a2640f995032e62 ("net: fix sock_wake_asy=
nc() rcu protection")
>>
>>
>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> compiler/cpufreq_governor/disk/filesize/fs2/fs/iterations/kconfig/nr_thr=
eads/rootfs/sync_method/tbox_group/test_size/testcase:
>>   gcc-4.9/performance/1BRD_48G/4M/nfsv4/xfs/1x/x86_64-rhel/64t/debian-x8=
6_64-2015-02-07.cgz/NoSync/lkp-hsx04/40G/fsmark
>>
>> commit:
>>   9cd3e072b0be17446e37d7414eac8a3499e0601e
>>   ceb5d58b217098a657f3850b7a2640f995032e62
>>
>> 9cd3e072b0be1744 ceb5d58b217098a657f3850b7a
>> ---------------- --------------------------
>>          %stddev     %change         %stddev
>
>> Thanks,
>> Ying Huang
>
> Hi Ying. I have no idea what these confusing numbers mean.
>
> Is your benchmark running faster (good thing) or slower (sorry, but we
> fixed a bug)

The benchmark running slower now.  This has big impact on performance,
maybe we can find some clue to optimize from the test result?  The
benchmark is about file IO performance on a loop back mounted NFS.

Best Regards,
Huang, Ying

--===============8337944977676653964==--


From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1751925AbbL1A61 (ORCPT <rfc822;w@1wt.eu>);
	Sun, 27 Dec 2015 19:58:27 -0500
Received: from mga02.intel.com ([134.134.136.20]:22484 "EHLO mga02.intel.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751164AbbL1A6T (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Sun, 27 Dec 2015 19:58:19 -0500
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.20,488,1444719600"; 
   d="scan'208";a="849241576"
From: "Huang\, Ying" <ying.huang@intel.com>
To: Eric Dumazet <edumazet@google.com>
Cc: <lkp@01.org>, LKML <linux-kernel@vger.kernel.org>,
        Dmitry Vyukov <dvyukov@google.com>,
        "David S. Miller" <davem@davemloft.net>
Subject: Re: [LKP] [net] ceb5d58b21: -69.2% fsmark.files_per_sec
References: <87si2sbdky.fsf@yhuang-dev.intel.com>
	<CANn89iJxF_qJwk0AmFVd57oexPzeB596KRL0r1HehP23JQM-dw@mail.gmail.com>
Date: Mon, 28 Dec 2015 08:58:17 +0800
In-Reply-To: <CANn89iJxF_qJwk0AmFVd57oexPzeB596KRL0r1HehP23JQM-dw@mail.gmail.com>
	(Eric Dumazet's message of "Thu, 24 Dec 2015 05:17:08 -0800")
Message-ID: <8760zj8k46.fsf@yhuang-dev.intel.com>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=ascii
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Eric Dumazet <edumazet@google.com> writes:

> On Wed, Dec 23, 2015 at 9:49 PM, kernel test robot
> <ying.huang@linux.intel.com> wrote:
>> FYI, we noticed the below changes on
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
>> commit ceb5d58b217098a657f3850b7a2640f995032e62 ("net: fix sock_wake_async() rcu protection")
>>
>>
>> =========================================================================================
>> compiler/cpufreq_governor/disk/filesize/fs2/fs/iterations/kconfig/nr_threads/rootfs/sync_method/tbox_group/test_size/testcase:
>>   gcc-4.9/performance/1BRD_48G/4M/nfsv4/xfs/1x/x86_64-rhel/64t/debian-x86_64-2015-02-07.cgz/NoSync/lkp-hsx04/40G/fsmark
>>
>> commit:
>>   9cd3e072b0be17446e37d7414eac8a3499e0601e
>>   ceb5d58b217098a657f3850b7a2640f995032e62
>>
>> 9cd3e072b0be1744 ceb5d58b217098a657f3850b7a
>> ---------------- --------------------------
>>          %stddev     %change         %stddev
>
>> Thanks,
>> Ying Huang
>
> Hi Ying. I have no idea what these confusing numbers mean.
>
> Is your benchmark running faster (good thing) or slower (sorry, but we
> fixed a bug)

The benchmark running slower now.  This has big impact on performance,
maybe we can find some clue to optimize from the test result?  The
benchmark is about file IO performance on a loop back mounted NFS.

Best Regards,
Huang, Ying