hama-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mikalai Parafeniuk (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HAMA-524) [GSoC 2012] Sparse Matrix-Vector multiplication (SpMV) on Hama
Date Sun, 15 Jul 2012 09:13:35 GMT

    [ https://issues.apache.org/jira/browse/HAMA-524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13414596#comment-13414596
] 

Mikalai Parafeniuk commented on HAMA-524:
-----------------------------------------

Hello everybody. I have published some basic working examples with test cases of SpMV in my
GitHub repositoty. But I have encountered some unclear problems. In debug mode in some configurations
(testSpMVSmall for LAMathTest) I noticed that different peers are from different sides of
barrier synchronization (peer.sync() method). For example I am getting NullPointerException
while trying to access array of local matrices
{code}
mLocal = mLocalArray[peerIndex];
{code}
Initialization of array is made in setup method
{code}
for (int i = 0; i < peerCount; i++) {
    mLocalArray[i] = strategy.getNewMatrixFormat();
    vLocalArray[i] = new SparseVector(m.getColumns());
}
{code}
I have placed additional peer.sync() in the end of setup method, but this doesn't helped.
Now I will be working to avoid in-memory storage of matrices (thanks Apurv for comment), but
I think this bug is crucial, and should be fixed as soon as possible. I think this occurs
because I made some wrong initialization. Could anyone review my current code, especially
test cases[1] and SpMV procedure[2] and help me to detect the problem?

[1]http://goo.gl/HPMv6
[2]http://goo.gl/gtwZ7
                
> [GSoC 2012] Sparse Matrix-Vector multiplication (SpMV) on Hama
> --------------------------------------------------------------
>
>                 Key: HAMA-524
>                 URL: https://issues.apache.org/jira/browse/HAMA-524
>             Project: Hama
>          Issue Type: New Feature
>          Components: bsp core, examples, math
>            Reporter: Edward J. Yoon
>            Assignee: Mikalai Parafeniuk
>              Labels: gsoc, gsoc2012, newbie
>
> Implement Efficient and Fast SpMV algorithm which can be widely used in scientiļ¬c computing,
financial modeling, information retrieval, and others, using Hama Bulk Synchronous Parallel
framework.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

       

Mime
View raw message