[
https://issues.apache.org/jira/browse/HAMA-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12675738#action_12675738
]
Edward J. Yoon commented on HAMA-127:
-------------------------------------
= Store Sparse Matrices =
Sparse Matrices in real applications are usually huge. Much space can be saved
if we only store non-zeros.
Space cost for all matrix entries = O(m*n)
Space cost for non-zeros entries = O(t)
Hbase stores the elements with their location (row, column, time). So it
provides a good representation.
(I guess we don't need to think more about structure/representation of sparse
matrix)
= Add/Mult =
Roughly, I think 'SIMDMultiplyMap' can be used for the sparse matrix
multiplication/addition if we only handle an non-zero element. Rests are almost
same.
> Sparse Vector/Matrix
> --------------------
>
> Key: HAMA-127
> URL: https://issues.apache.org/jira/browse/HAMA-127
> Project: Hama
> Issue Type: New Feature
> Components: implementation
> Affects Versions: 0.1.0
> Reporter: Edward J. Yoon
> Fix For: 0.1.0
>
>
> Let's add sparse vector/matrix feature to 0.1
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
