[PATCH v4] Improve the performance of --num-threads -d 31

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



>> Hello,
>>
>> This is just a quick note to inform you.
>> I measured the memory consumption with -d31 by VmHWM in
>> /proc/PID/status and compared them between v3 and v4 since
>> Minfei said the problem only occurs in v4.
>>
>>              |          VmHWM[kB]
>> num-thread  |      v3            v4
>> ------------+--------------------------
>>       1      |    20,516        20,516
>>       2      |    20,624        20,628
>>       4      |    20,832        20,832
>>       8      |    21,292        21,288
>>      16      |    22,240        22,236
>>      32      |    24,096        24,100
>>      64      |    27,900        27,888
>>
>> According to this result, the problem we face seems not just
>> any lack of memory issue.
>>
>
>Yes, I had realized it, for there isn't much difference between v3 and v4.
>And it is hardly to some further investigation, until get Minfei's result.
>
>BTW, can you reproduce the bug?

Unfortunately, I can't reproduce it yet.
I'm also waiting for Minfei's result.


Thanks,
Atsushi Kumagai

>> BTW, the memory consumption increases depending on num-thread,
>> I think it should be considered in the calculate_cyclic_buffer_size().
>>
>
>I will think about it.
>
>--
>Thanks
>Zhou
>
>>
>> Thanks,
>> Atsushi Kumagai
>>
>> diff --git a/makedumpfile.c b/makedumpfile.c
>> index 4075f3e..d5626f9 100644
>> --- a/makedumpfile.c
>> +++ b/makedumpfile.c
>> @@ -44,6 +44,14 @@ extern int find_vmemmap();
>>
>>   char filename_stdout[] = FILENAME_STDOUT;
>>
>> +void
>> +print_VmHWM(void)
>> +{
>> +       char command[64];
>> +       sprintf(command, "grep VmHWM /proc/%d/status", getpid());
>> +       system(command);
>> +}
>> +
>>   /* Cache statistics */
>>   static unsigned long long      cache_hit;
>>   static unsigned long long      cache_miss;
>> @@ -11185,5 +11193,7 @@ out:
>>          }
>>          free_elf_info();
>>
>> +       print_VmHWM();
>> +
>>          return retcd;
>>   }
>>
>>
>>> Hi, Zhou.
>>>
>>> I'm on holiday now, you can ask other people to help test, if necessary.
>>>
>>> Thanks
>>> Minfei
>>>
>>>> ? 2016?3?24??12:29?Zhou, Wenjian/??? <zhouwj-fnst at cn.fujitsu.com> ???
>>>>
>>>> Hello Minfei,
>>>>
>>>> How do these two patches work?
>>>>
>>>> --
>>>> Thanks
>>>> Zhou
>>>>
>>>>> On 03/18/2016 01:48 PM, "Zhou, Wenjian/???" wrote:
>>>>>> On 03/18/2016 12:16 PM, Minfei Huang wrote:
>>>>>>> On 03/18/16 at 10:46am, "Zhou, Wenjian/???" wrote:
>>>>>>> Hello Minfei,
>>>>>>>
>>>>>>> Since I can't produce the bug, I reviewed the patch and wrote an increment patch.
>>>>>>> Though there are some bugs in the increment patch,
>>>>>>> I wonder if the previous bug still exists with this patch.
>>>>>>> Could you help me confirm it?
>>>>>>
>>>>>> Ok. I will help verify this increasing patch.
>>>>>
>>>>> Thank you very much.
>>>>>
>>>>>>>
>>>>>>> And I have another question.
>>>>>>> Did it only occur in patch v4?
>>>>>>
>>>>>> This issue doesn't exist in v3. I have pasted the test result with
>>>>>> --num-thread 32 in that thread.
>>>>>>
>>>>>> applied makedumpfile with option -d 31 --num-threads 32
>>>>>> real    3m3.533s
>>>>>
>>>>> Oh, then the patch in the previous mail may not work.
>>>>>
>>>>> I'm appreciated if you can also test the patch in this letter.
>>>>>
>>>>> I introduced semaphore to fix the bug in the v3.
>>>>> So I want to know if it is this which affects the result.
>>>>> The attached patch is based on v4, used to remove semaohore.
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> kexec mailing list
>>>>> kexec at lists.infradead.org
>>>>> http://lists.infradead.org/mailman/listinfo/kexec
>>>>
>>>>
>>> _______________________________________________
>>> kexec mailing list
>>> kexec at lists.infradead.org
>>> http://lists.infradead.org/mailman/listinfo/kexec
>>
>>
>



[Index of Archives]     [LM Sensors]     [Linux Sound]     [ALSA Users]     [ALSA Devel]     [Linux Audio Users]     [Linux Media]     [Kernel]     [Gimp]     [Yosemite News]     [Linux Media]

  Powered by Linux