systemml-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matthias Boehm (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SYSTEMML-1811) Can we Implement X%*%t(X) in a better way?
Date Wed, 26 Jul 2017 21:28:00 GMT

    [ https://issues.apache.org/jira/browse/SYSTEMML-1811?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16102314#comment-16102314
] 

Matthias Boehm commented on SYSTEMML-1811:
------------------------------------------

For all backends (CP, SPARK, MR), we actually have physical operators, called tsmm (transpose-self
matrix multiplication) for t(X)%*%X and X%*%t(X). At block level there are also special implementations
for this operation:

https://github.com/apache/systemml/blob/master/src/main/java/org/apache/sysml/runtime/matrix/data/LibMatrixMult.java#L1648

> Can we Implement X%*%t(X) in a better way?
> ------------------------------------------
>
>                 Key: SYSTEMML-1811
>                 URL: https://issues.apache.org/jira/browse/SYSTEMML-1811
>             Project: SystemML
>          Issue Type: Improvement
>            Reporter: Janardhan
>
> A matrix multiplied by its self transpose is a frequent occurrence in many algorithms
( a lot of them). There is definitely a way to take into consideration the special properties
of this matrix operation.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message