[
https://issues.apache.org/jira/browse/SPARK-16696?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Weichen Xu updated SPARK-16696:
-------------------------------
Description:
Unused broadcast variables should call destroy() instead of unpersist() so that
the memory can released in time, even in driver-side.
currently, several algorithm in ML, such as KMeans, Word2Vec, has this problem.
was:
Unused broadcast variables should call destroy() instead of unpersist() so that
the memory can released in time, even in driver-side.
currently, several algorithm in ML, such as KMeans, LDA, Word2Vec, has this
problem. Several places unused broadcast vars even neither call unpersit() nor
call destroy()
> unused broadcast variables should call destroy instead of unpersist
> -------------------------------------------------------------------
>
> Key: SPARK-16696
> URL: https://issues.apache.org/jira/browse/SPARK-16696
> Project: Spark
> Issue Type: Bug
> Components: ML, MLlib
> Affects Versions: 2.0.1, 2.1.0
> Reporter: Weichen Xu
> Original Estimate: 1m
> Remaining Estimate: 1m
>
> Unused broadcast variables should call destroy() instead of unpersist() so
> that the memory can released in time, even in driver-side.
> currently, several algorithm in ML, such as KMeans, Word2Vec, has this
> problem.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
