[jira] [Commented] (SPARK-19530) Use guava weigher for code cache eviction

["Apache Spark (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-19530?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15859157#comment-15859157 ] 

Apache Spark commented on SPARK-19530:
--------------------------------------

User 'viirya' has created a pull request for this issue:
https://github.com/apache/spark/pull/16865

> Use guava weigher for code cache eviction
> -----------------------------------------
>
>                 Key: SPARK-19530
>                 URL: https://issues.apache.org/jira/browse/SPARK-19530
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 2.1.0
>            Reporter: Liang-Chi Hsieh
>
> We use guava cache to cache compiled codes for codegen. Currently we use number of entries (100 as maximum now) in the cache to determine when to evict older entries.
> However, the number of entries can't respond well to actually memory usage of cache entries. As we heavily rely codegen now and the generated codes can be large, we shouldn't use maximum of entries for code cache entry eviction.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org