>On 06/26/2015 03:07 PM, Atsushi Kumagai wrote:
>> Hello Zhou,
>>
>>> >Hello Atsushi Kumagai,
>>> >
>>> >I test this patch set in several machines and the following is the benchmark.
>> Thanks for your report, it looks good as before.
>> I also did simple test on kernel 3.19 on a 5GB Virtual Machine,
>> but I can't get such good result as below:
>>
>> / # time makedumpfile -c --num-threads 0 /proc/vmcore /mnt/dumpfile
>> Copying data : [100.0 %] /
>>
>> The dumpfile is saved to /mnt/dumpfile.
>>
>> makedumpfile Completed.
>> real 0m 44.40s
>> user 0m 43.62s
>> sys 0m 0.69s
>> / #
>>
>> / # time makedumpfile -c --num-threads 4 /proc/vmcore /mnt/dumpfile
>> Copying data : [100.0 %] -
>> Copying data : [100.0 %] /
>>
>> The dumpfile is saved to /mnt/dumpfile.
>>
>> makedumpfile Completed.
>> real 5m 29.54s
>> user 6m 8.18s
>> sys 16m 33.25s
>> / #
>>
>>
>> There is a big performance degradation.
>> Do you have any ideas why this happens ?
>
>Does your Virtual Machine have more than 2 processors?
>If so, check the value:nr_cpus in KDUMP_COMMANDLINE_APPEND of /etc/sysconfig/kdump.
>It should be also set to the number larger than 2.
>
>To "--num-threads 4", the most suitable number of processors is 5.
I attached 5 processors to the VM and I confirmed that all threads
consumed full cpu time by top(1) on the host:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17614 qemu 20 0 5792m 4.9g 5652 R 435.1 72.8 29:02.17 qemu-kvm
So I think the performance must be improved...
Thanks
Atsushi Kumagai
