[jira] Issue Comment Edited: (HAMA-129) Improving speed of matrix multiplication

["Edward J. Yoon (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HAMA-129?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12657671#action_12657671 ] 

udanax edited comment on HAMA-129 at 12/17/08 8:48 PM:
---------------------------------------------------------------

Each submatrix should be as large as possible. However, large cell causes RetriesExhaustedException during HTable.commit.
>> Looks like HADOOP-4802 (in hbase, its part of HBASE-900).  Big cells can trigger OOME. 

I would like to update hadoop/hbase. I think we can't release with hadoop/hbase 0.18.x

      was (Author: udanax):
    Each submatrix should be as large as possible. However, large cell causes RetriesExhaustedException during HTable.commit.
Looks like HADOOP-4802 (in hbase, its part of HBASE-900).  Big cells can trigger OOME. 

Let's update hadoop/hbase.
  
> Improving speed of matrix multiplication
> ----------------------------------------
>
>                 Key: HAMA-129
>                 URL: https://issues.apache.org/jira/browse/HAMA-129
>             Project: Hama
>          Issue Type: Improvement
>          Components: implementation
>    Affects Versions: 0.1.0
>            Reporter: Edward J. Yoon
>            Assignee: Edward J. Yoon
>             Fix For: 0.1.0
>
>         Attachments: HAMA-129.patch, HAMA-129_v01.patch, HAMA-129_v02.patch, HAMA-129_v03.patch
>
>
> We use a blockinputformat (row based partitioning). but it causes "java.lang.OutOfMemoryError: Java heap space" when size is big.
> I'll write the formulas and partition them as PIEstimator.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.