[
https://issues.apache.org/jira/browse/HAMA-999?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
JongYoon Lim updated HAMA-999:
------------------------------
Description:
I found that *SuperstepPiEstimator* example sometimes gives wrong result when
call *peer.getNumCurrentMessages()*. And that was because of wrong *size()* of
*MemoryQueue*. When I printed out sizes of queue from the example, sometimes it
said,
{noformat}
bundle size: 20, numOfMsg: 19, deque size: 0
{noformat}
I think *deque*, *bundles* and *numOfMsg* of *MemoryQueue* should be properly
synchronized to get correct result.
was:
I found that *SuperstepPiEstimator* example sometimes gives wrong result when
call *peer.getCurrentMessage()*. And that was because of wrong *size()* of
*MemoryQueue*. When I printed out sizes of queue from the example, sometimes it
said,
{noformat}
bundle size: 20, numOfMsg: 19, deque size: 0
{noformat}
I think *deque*, *bundles* and *numOfMsg* of *MemoryQueue* should be properly
synchronized to get correct result.
> Wrong size of MemoryQueue
> -------------------------
>
> Key: HAMA-999
> URL: https://issues.apache.org/jira/browse/HAMA-999
> Project: Hama
> Issue Type: Bug
> Components: bsp core
> Reporter: JongYoon Lim
>
> I found that *SuperstepPiEstimator* example sometimes gives wrong result when
> call *peer.getNumCurrentMessages()*. And that was because of wrong *size()*
> of *MemoryQueue*. When I printed out sizes of queue from the example,
> sometimes it said,
> {noformat}
> bundle size: 20, numOfMsg: 19, deque size: 0
> {noformat}
> I think *deque*, *bundles* and *numOfMsg* of *MemoryQueue* should be properly
> synchronized to get correct result.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
