[
https://issues.apache.org/jira/browse/HAMA-742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13596442#comment-13596442
]
Thomas Jungblut commented on HAMA-742:
--------------------------------------
Seems we are talking about different buffers here ;-) I was talking about the
overall size (so the sum of your intermediate buffers). In MPI this would be
64m/512 task on average = 128k for each peer so I guess we can lift it a bit
without overflowing memory too much.
Performance improvement is always relative to your network bandwidth, I guess
you were simulating with virtual boxes, right?
bq.Unfortunately, my code is across projects with some passing and some failing
unit tests. I shall upload few building blocks by this weekend.
Cool, let me know when something is there. I will have a deeper look.
> Implement of Hama RPC
> ----------------------
>
> Key: HAMA-742
> URL: https://issues.apache.org/jira/browse/HAMA-742
> Project: Hama
> Issue Type: Sub-task
> Reporter: Edward J. Yoon
> Fix For: 0.6.1
>
>
> To solve HDFS 2.0 compatibility issue, we have to change a lot of codes for
> Hadoop 2.0 RPC, moreover, yarn RPC doesn't support asynchronous call directly.
> Ultimately, we can pursue the performance and integrate more easily with
> hadoop multi-versions by having our own RPC.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
