ignite-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alisher Alimov <alimovalis...@gmail.com>
Subject Re: Performance question
Date Tue, 29 Nov 2016 07:11:33 GMT
> int[] partitions =
> this.ignite.affinity("test.cache").primaryPartitions(this.ignite.cluster().node());
> 
> cluster().node() returns a random node from the group, it should be
> localNode() instead if I understand your approach correctly.

Because I have only one node in cluster, current method will return instance of local node.

> I would also try to manually start threads to make sure that parallelization
> works properly. I'm not sure how it works if you use streams.

Parallel stream will use fork join model and use more than 1 CPU 

Here is screenshot from profiler 



I have tried to use Executor but I didn't achieve more performance

With best regards
Alisher Alimov
alimovalisher@gmail.com <mailto:alimovalisher@gmail.com>




> On 29 Nov 2016, at 00:52, vkulichenko <valentin.kulichenko@gmail.com <mailto:valentin.kulichenko@gmail.com>>
wrote:
> 
> Alisher,
> 
> This line does not look correct to me:
> 
> int[] partitions =
> this.ignite.affinity("test.cache").primaryPartitions(this.ignite.cluster().node());
> 
> cluster().node() returns a random node from the group, it should be
> localNode() instead if I understand your approach correctly.
> 
> I would also try to manually start threads to make sure that parallelization
> works properly. I'm not sure how it works if you use streams.
> 
> -Val
> 
> 
> 
> --
> View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Performance-question-tp9218p9239.html
<http://apache-ignite-users.70518.x6.nabble.com/Performance-question-tp9218p9239.html>
> Sent from the Apache Ignite Users mailing list archive at Nabble.com <http://nabble.com/>.


Mime
View raw message