Hi Semih,
interesting email. I'm probably not getting your technique right, but
why wouldn't it be possible to compute the master.compute() inside of
an aggregator?
Not only it *should* be possible, but as aggregators are computed both
on workers AND on the master, you should have a faster
[
https://issues.apache.org/jira/browse/GIRAPH-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13189574#comment-13189574
]
Avery Ching commented on GIRAPH-127:
I think this functionality is very useful and
