[
https://issues.apache.org/jira/browse/MADLIB-1067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16215817#comment-16215817
]
Frank McQuillan commented on MADLIB-1067:
-----------------------------------------
Updated conversation
http://mail-archives.apache.org/mod_mbox/madlib-dev/201710.mbox/%3CCAKG0DM1b5%3D4Sx3u80wq5-F3WagzfM%2BtBWBk-%2BYyREHc_ZL6bbA%40mail.gmail.com%3E
which describes other modules that could benefit as well.
As suggested in the thread, this might be a good 2.0 feature.
> Use Eigen in the implementation of array_dot
> --------------------------------------------
>
> Key: MADLIB-1067
> URL: https://issues.apache.org/jira/browse/MADLIB-1067
> Project: Apache MADlib
> Issue Type: Improvement
> Components: Module: Linear-Algebra Functions
> Reporter: Mauricio Scheffer
> Fix For: v2.0
>
>
> Currently (MADlib 1.9.1) cosine_similarity has better performance than
> array_dot.
> It seems that array_dot should be using the dot product implementation from
> Eigen, like cosine_similarity.
> See
> http://mail-archives.apache.org/mod_mbox/incubator-madlib-user/201702.mbox/%3CCAAKWcNHoxdVdEjRsAOjC9Zukjo2e5cLBP2XPFFC6f8XiXSMykA%40mail.gmail.com%3E
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
