[
https://issues.apache.org/jira/browse/MAHOUT-308?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12880115#action_12880115
]
Jake Mannix commented on MAHOUT-308:
------------------------------------
Just a quick comment (still settling in GMT+1, actually!): I think this
approach is the right one, and cleaner than my off-the-cuff approach, API-wise.
(I need to run it, of course! Will try that in the next few days).
The inefficiency of the desiredRank loops over the basis set is something that
can be solved in a further patch, yes.
> Improve Lanczos to handle extremely large feature sets (without hashing)
> ------------------------------------------------------------------------
>
> Key: MAHOUT-308
> URL: https://issues.apache.org/jira/browse/MAHOUT-308
> Project: Mahout
> Issue Type: Improvement
> Components: Math
> Affects Versions: 0.3
> Environment: all
> Reporter: Jake Mannix
> Assignee: Jake Mannix
> Fix For: 0.4
>
> Attachments: MAHOUT-308.patch
>
>
> DistributedLanczosSolver currently keeps all Lanczos vectors in memory on the
> driver (client) computer while Hadoop is iterating. The memory requirements
> of this is (desiredRank) * (numColumnsOfInput) * 8bytes, which for
> desiredRank = a few hundred, starts to cap out usefulness at
> some-small-number * millions of columns for most commodity hardware.
> The solution (without doing stochastic decomposition) is to persist the
> Lanczos basis to disk, except for the most recent two vectors. Some care
> must be taken in the "orthogonalizeAgainstBasis()" method call, which uses
> the entire basis. This part would be slower this way.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
