[
https://issues.apache.org/jira/browse/MAHOUT-1190?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13628853#comment-13628853
]
Sebastian Schelter edited comment on MAHOUT-1190 at 4/11/13 11:48 AM:
----------------------------------------------------------------------
I took a look in the code and the situation is even worse, as DenseVector
overrides assign and only has a special handling for PlusMult (not Plus!)
I think we should extend the interfaces, so that you can ask a function whether
f(0,x) = 0 and f(x,0) = x holds. I think we should keep definitely keep SASV
and make sure that all vectors behave best as possible.
was (Author: ssc):
I took a look in the code and the situation is even worse, as DenseVector
overrides assign and only has a special handling for PlusMult (not Plus!)
I think we should extend the interfaces, so that you can ask a function whether
f(0,x) = 0 and f(x,0) holds. I think we should keep definitely keep SASV and
make sure that all vectors behave best as possible.
> SequentialAccessSparseVector function assignment is very slow
> -------------------------------------------------------------
>
> Key: MAHOUT-1190
> URL: https://issues.apache.org/jira/browse/MAHOUT-1190
> Project: Mahout
> Issue Type: Bug
> Reporter: Dan Filimon
>
> Currently when calling .assign() on a SASV with another vector and a custom
> function, it will iterate through it and assign every single entry while also
> referring it by index.
> This makes the process *hugely* expensive. (on a run of BallKMeans on the 20
> newsgroups data set, profiling reveals that 92% of the runtime was spent
> updating assigning the vectors).
> Here's a prototype patch:
> https://github.com/dfilimon/mahout/commit/63998d82bb750150a6ae09052dadf6c326c62d3d
--
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
