Re: [LKP] [lkp-developer] [sched/fair] 4e5160766f: +149% ftq.noise.50% regression

[Vincent Guittot]

Hi Ying,

On 21 February 2017 at 03:40, Huang, Ying <ying.huang@xxxxxxxxx> wrote:
> Hi, Vincent,
>
> Vincent Guittot <vincent.guittot@xxxxxxxxxx> writes:
>

[snip]

>>>
>>> Here is the test result,
>>>
>>> =========================================================================================
>>> compiler/cpufreq_governor/freq/kconfig/nr_task/rootfs/samples/tbox_group/test/testcase:
>>>   gcc-6/powersave/20/x86_64-rhel-7.2/100%/debian-x86_64-2016-08-31.cgz/6000ss/lkp-hsw-d01/cache/ftq
>>>
>>> commit:
>>>   4e5160766fcc9f41bbd38bac11f92dce993644aa: first bad commit
>>>   09a43ace1f986b003c118fdf6ddf1fd685692d49: parent of first bad commit
>>>   b524060933c546fd2410c5a09360ba23a0fef846: with fix patch above
>>>
>>> 4e5160766fcc9f41 09a43ace1f986b003c118fdf6d b524060933c546fd2410c5a093
>>> ---------------- -------------------------- --------------------------
>>>          %stddev     %change         %stddev     %change         %stddev
>>>              \          |                \          |                \
>>>       3463 Â 10%     -61.4%       1335 Â 17%      -3.0%       3359 Â  2%  ftq.noise.50%
>>>       1116 Â 23%     -73.7%     293.90 Â 30%     -23.8%     850.69 Â 17%  ftq.noise.75%
>>
>> To be honest, I was expecting at least the same level of improvement
>> as the previous patch if not better but i was not expecting worse
>> results
>
> What's your next plan for this regression?  At least your previous patch
> could recover part of it.

I haven't been able to find better fix than the previous patch so i'm
going to send a clean version with proper commit message

Regards,
Vincent

>
> Best Regards,
> Huang, Ying