Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   "Huang, Ying" <ying.huang@intel.com>
To:     Shakeel Butt <shakeelb@google.com>
Cc:     kernel test robot <yujie.liu@intel.com>, oe-lkp@lists.linux.dev,
        lkp@intel.com, Andrew Morton <akpm@linux-foundation.org>,
        Marek Szyprowski <m.szyprowski@samsung.com>,
        Linux Memory Management List <linux-mm@kvack.org>,
        linux-kernel@vger.kernel.org, feng.tang@intel.com,
        zhengjun.xing@linux.intel.com, fengwei.yin@intel.com
Subject: Re: [linux-next:master] [mm] f1a7941243: unixbench.score -5.1%
 regression
References: <202212051324.182a6363-yujie.liu@intel.com>
        <CALvZod4CD+O7-ynGAHU-6vxE6CbSmuQei1=SVJsx0zFfQdmV2g@mail.gmail.com>
Date:   Tue, 06 Dec 2022 10:41:00 +0800
In-Reply-To: <CALvZod4CD+O7-ynGAHU-6vxE6CbSmuQei1=SVJsx0zFfQdmV2g@mail.gmail.com>
        (Shakeel Butt's message of "Mon, 5 Dec 2022 09:18:28 -0800")
Message-ID: <875yepfe4j.fsf@yhuang6-desk2.ccr.corp.intel.com>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Shakeel Butt <shakeelb@google.com> writes:

> On Sun, Dec 4, 2022 at 9:56 PM kernel test robot <yujie.liu@intel.com> wr=
ote:
>>
>> Greeting,
>>
>> FYI, we noticed a -5.1% regression of unixbench.score due to commit:
>>
> [...]
>> 9cd6ffa60256e931 f1a7941243c102a44e8847e3b94
>> ---------------- ---------------------------
>>          %stddev     %change         %stddev
>>              \          |                \
>>       7917            -5.1%       7509        unixbench.score
>
> What is unixbench.score?

Should be benchmark throughput.

>>      10485           -12.1%       9216        unixbench.time.maximum_res=
ident_set_size

This should reflect accuracy change of per_cpu_counter.

>>   37236706            -5.1%   35324104        unixbench.time.minor_page_=
faults

The reduction is same as benchmark score.  So I think this reflect the
nature of time-bound testing (instead of workload-bound).

> For above two, is negative change good or bad?
>
>>       0.98 =C4=85 20%      +0.7        1.64 =C4=85 38%  perf-profile.cal=
ltrace.cycles-pp.link_path_walk.path_openat.do_filp_open.do_sys_openat2.__x=
64_sys_openat
>>       2.12 =C4=85 19%      +0.8        2.96 =C4=85 13%  perf-profile.cal=
ltrace.cycles-pp.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_=
page_fault
>>       2.35 =C4=85 13%      +0.9        3.28 =C4=85 13%  perf-profile.cal=
ltrace.cycles-pp.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_p=
age_fault.asm_exc_page_fault
>>       0.14 =C4=85 74%      +0.4        0.55 =C4=85 32%  perf-profile.chi=
ldren.cycles-pp.do_task_dead
>>       0.04 =C4=85223%      +0.4        0.47 =C4=85 49%  perf-profile.chi=
ldren.cycles-pp.__mmdrop
>
> Also how should I interpret the above perf-profiles?

It appears that the changes of handle_mm_fault() and __mmdrop() are
related to the code of the commit?  That is, for this specific workloads
(not so unpractical), the operations become slower?

Best Regards,
Huang, Ying