[
https://issues.apache.org/jira/browse/MAHOUT-1736?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14599903#comment-14599903
]
ASF GitHub Bot commented on MAHOUT-1736:
----------------------------------------
Github user avati commented on the pull request:
https://github.com/apache/mahout/pull/143#issuecomment-114971698
@dlyubimov H2OBlockMatrix should work like any other in-core matrix in the
backend. In fact its lifecycle is valid only in the backend (you cannot spin-up
a H2OBlockMatrix from scratch in the "frontend").
Yes, you are right that we are not running all the in-core matrix tests. We
will have to redesign all the tests to be run in the context of a MapBlock or
allreduceBlock closure. That's the only way to create a H2OBlockMatrix.
> H20 implementation for allreduceBlock
> -------------------------------------
>
> Key: MAHOUT-1736
> URL: https://issues.apache.org/jira/browse/MAHOUT-1736
> Project: Mahout
> Issue Type: Improvement
> Affects Versions: 0.10.1
> Reporter: Dmitriy Lyubimov
> Assignee: Anand Avati
> Fix For: 0.10.2
>
>
> allreduce aggregator on matrix blocks:
> {code}
> override def allreduceBlock[K: ClassTag](drm: CheckpointedDrm[K], bmf:
> BlockMapFunc2[K], rf: BlockReduceFunc)
> : Matrix = ???
> {code}
> This is pretty important for custom aggregated statistics
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
