You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ignite.apache.org by Anton Vinogradov <av...@apache.org> on 2018/08/03 10:33:09 UTC
[BENCHMARK REQUEST] IGNITE-8446
Dmitry,
AFAIK, you're able to perform benchmarking of critical code.
Could you please check https://issues.apache.org/jira/browse/IGNITE-8446?
We have to check
https://github.com/apache/ignite/tree/ignite-8446
vs
https://github.com/apache/ignite/tree/ignite-8446-master
Only one benchmark is required: tx throughput
Re: [BENCHMARK REQUEST] IGNITE-8446
Posted by Anton Vinogradov <av...@apache.org>.
Great news.
Ilya, thank you!
пт, 3 авг. 2018 г. в 17:45, Ilya Suntsov <is...@gridgain.com>:
> Anton,
>
> I've run benchmarks with the following parameters:
>
> - 4 servers, 8 clients
> - primary/full sync
> - in-memory
>
> Benchmarks show approximately the same numbers.
>
> Primary sync:
> ignite-8446 ignite-8446-master delta
> tx-optim-repRead-put-get 36119 36241 -0.34%
> tx-pessim-repRead-put-get 16747 16886 -0.83%
> tx-putAll 702 706 -0.57%
>
> Full sync:
> ignite-8446 ignite-8446-master delta
> tx-optim-repRead-put-get 35967 36386 -1.16%
> tx-pessim-repRead-put-get 14415 14513 -0.68%
> tx-putAll 693 695 -0.29%
>
> 2018-08-03 15:53 GMT+03:00 Anton Vinogradov <av...@apache.org>:
>
>> - Benchmark should check how many tx can be completed in second (please
>> use same as you use to check the release)
>> - in-memory
>> - default
>> - same you use at pre-release check.
>>
>> пт, 3 авг. 2018 г. в 15:38, Ilya Suntsov <is...@gridgain.com>:
>>
>>> Anton,
>>>
>>> Please provide more details:
>>>
>>> - Benchmark class name
>>> - PDS/in-memory
>>> - sync mode
>>> - cluster size
>>>
>>> ?
>>>
>>> 2018-08-03 13:52 GMT+03:00 Dmitriy Pavlov <dp...@gmail.com>:
>>>
>>>> Hi Anton,
>>>>
>>>> Thank you for bringing this question here. Actually I'm not right
>>>> person who can do it, but hopefully Ilya can assist.
>>>>
>>>> Hi Ilya,
>>>>
>>>> could you please run required benchmark?
>>>>
>>>> Sincerely,
>>>> Dmitriy Pavlov
>>>>
>>>> пт, 3 авг. 2018 г. в 13:33, Anton Vinogradov <av...@apache.org>:
>>>>
>>>>> Dmitry,
>>>>>
>>>>> AFAIK, you're able to perform benchmarking of critical code.
>>>>> Could you please check
>>>>> https://issues.apache.org/jira/browse/IGNITE-8446?
>>>>>
>>>>> We have to check
>>>>> https://github.com/apache/ignite/tree/ignite-8446
>>>>> vs
>>>>> https://github.com/apache/ignite/tree/ignite-8446-master
>>>>>
>>>>> Only one benchmark is required: tx throughput
>>>>>
>>>>
>>>
>>>
>>> --
>>> Best Regards,
>>> Ilya Suntsov
>>> email: isuntsov@gridgain.com
>>> *GridGain Systems*
>>> www.gridgain.com
>>>
>>
>
>
> --
> Best Regards,
> Ilya Suntsov
> email: isuntsov@gridgain.com
> *GridGain Systems*
> www.gridgain.com
>
Re: [BENCHMARK REQUEST] IGNITE-8446
Posted by Ilya Suntsov <is...@gridgain.com>.
Anton,
I've run benchmarks with the following parameters:
- 4 servers, 8 clients
- primary/full sync
- in-memory
Benchmarks show approximately the same numbers.
Primary sync:
ignite-8446 ignite-8446-master delta
tx-optim-repRead-put-get 36119 36241 -0.34%
tx-pessim-repRead-put-get 16747 16886 -0.83%
tx-putAll 702 706 -0.57%
Full sync:
ignite-8446 ignite-8446-master delta
tx-optim-repRead-put-get 35967 36386 -1.16%
tx-pessim-repRead-put-get 14415 14513 -0.68%
tx-putAll 693 695 -0.29%
2018-08-03 15:53 GMT+03:00 Anton Vinogradov <av...@apache.org>:
> - Benchmark should check how many tx can be completed in second (please
> use same as you use to check the release)
> - in-memory
> - default
> - same you use at pre-release check.
>
> пт, 3 авг. 2018 г. в 15:38, Ilya Suntsov <is...@gridgain.com>:
>
>> Anton,
>>
>> Please provide more details:
>>
>> - Benchmark class name
>> - PDS/in-memory
>> - sync mode
>> - cluster size
>>
>> ?
>>
>> 2018-08-03 13:52 GMT+03:00 Dmitriy Pavlov <dp...@gmail.com>:
>>
>>> Hi Anton,
>>>
>>> Thank you for bringing this question here. Actually I'm not right person
>>> who can do it, but hopefully Ilya can assist.
>>>
>>> Hi Ilya,
>>>
>>> could you please run required benchmark?
>>>
>>> Sincerely,
>>> Dmitriy Pavlov
>>>
>>> пт, 3 авг. 2018 г. в 13:33, Anton Vinogradov <av...@apache.org>:
>>>
>>>> Dmitry,
>>>>
>>>> AFAIK, you're able to perform benchmarking of critical code.
>>>> Could you please check https://issues.apache.
>>>> org/jira/browse/IGNITE-8446?
>>>>
>>>> We have to check
>>>> https://github.com/apache/ignite/tree/ignite-8446
>>>> vs
>>>> https://github.com/apache/ignite/tree/ignite-8446-master
>>>>
>>>> Only one benchmark is required: tx throughput
>>>>
>>>
>>
>>
>> --
>> Best Regards,
>> Ilya Suntsov
>> email: isuntsov@gridgain.com
>> *GridGain Systems*
>> www.gridgain.com
>>
>
--
Best Regards,
Ilya Suntsov
email: isuntsov@gridgain.com
*GridGain Systems*
www.gridgain.com
Re: [BENCHMARK REQUEST] IGNITE-8446
Posted by Anton Vinogradov <av...@apache.org>.
- Benchmark should check how many tx can be completed in second (please use
same as you use to check the release)
- in-memory
- default
- same you use at pre-release check.
пт, 3 авг. 2018 г. в 15:38, Ilya Suntsov <is...@gridgain.com>:
> Anton,
>
> Please provide more details:
>
> - Benchmark class name
> - PDS/in-memory
> - sync mode
> - cluster size
>
> ?
>
> 2018-08-03 13:52 GMT+03:00 Dmitriy Pavlov <dp...@gmail.com>:
>
>> Hi Anton,
>>
>> Thank you for bringing this question here. Actually I'm not right person
>> who can do it, but hopefully Ilya can assist.
>>
>> Hi Ilya,
>>
>> could you please run required benchmark?
>>
>> Sincerely,
>> Dmitriy Pavlov
>>
>> пт, 3 авг. 2018 г. в 13:33, Anton Vinogradov <av...@apache.org>:
>>
>>> Dmitry,
>>>
>>> AFAIK, you're able to perform benchmarking of critical code.
>>> Could you please check https://issues.apache.org/jira/browse/IGNITE-8446
>>> ?
>>>
>>> We have to check
>>> https://github.com/apache/ignite/tree/ignite-8446
>>> vs
>>> https://github.com/apache/ignite/tree/ignite-8446-master
>>>
>>> Only one benchmark is required: tx throughput
>>>
>>
>
>
> --
> Best Regards,
> Ilya Suntsov
> email: isuntsov@gridgain.com
> *GridGain Systems*
> www.gridgain.com
>
Re: [BENCHMARK REQUEST] IGNITE-8446
Posted by Ilya Suntsov <is...@gridgain.com>.
Anton,
Please provide more details:
- Benchmark class name
- PDS/in-memory
- sync mode
- cluster size
?
2018-08-03 13:52 GMT+03:00 Dmitriy Pavlov <dp...@gmail.com>:
> Hi Anton,
>
> Thank you for bringing this question here. Actually I'm not right person
> who can do it, but hopefully Ilya can assist.
>
> Hi Ilya,
>
> could you please run required benchmark?
>
> Sincerely,
> Dmitriy Pavlov
>
> пт, 3 авг. 2018 г. в 13:33, Anton Vinogradov <av...@apache.org>:
>
>> Dmitry,
>>
>> AFAIK, you're able to perform benchmarking of critical code.
>> Could you please check https://issues.apache.org/jira/browse/IGNITE-8446?
>>
>> We have to check
>> https://github.com/apache/ignite/tree/ignite-8446
>> vs
>> https://github.com/apache/ignite/tree/ignite-8446-master
>>
>> Only one benchmark is required: tx throughput
>>
>
--
Best Regards,
Ilya Suntsov
email: isuntsov@gridgain.com
*GridGain Systems*
www.gridgain.com
Re: [BENCHMARK REQUEST] IGNITE-8446
Posted by Dmitriy Pavlov <dp...@gmail.com>.
Hi Anton,
Thank you for bringing this question here. Actually I'm not right person
who can do it, but hopefully Ilya can assist.
Hi Ilya,
could you please run required benchmark?
Sincerely,
Dmitriy Pavlov
пт, 3 авг. 2018 г. в 13:33, Anton Vinogradov <av...@apache.org>:
> Dmitry,
>
> AFAIK, you're able to perform benchmarking of critical code.
> Could you please check https://issues.apache.org/jira/browse/IGNITE-8446?
>
> We have to check
> https://github.com/apache/ignite/tree/ignite-8446
> vs
> https://github.com/apache/ignite/tree/ignite-8446-master
>
> Only one benchmark is required: tx throughput
>