[ https://issues.apache.org/jira/browse/SPARK-26337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16717665#comment-16717665 ]
ASF GitHub Bot commented on SPARK-26337: ---------------------------------------- kiszk commented on issue #23284: [SPARK-26337][SQL][TEST] Add benchmark for LongToUnsafeRowMap URL: https://github.com/apache/spark/pull/23284#issuecomment-446299651 To be honest, I cannot understand why the original performance degradation occurred. I think that read/write of `long` value does not require any sychronization or memory barrier without declaring `volatile`. At this [PR](https://github.com/apache/spark/pull/23204/files), `numKeyLookups` and `numProbes` are non-volatile long variables. I confirmed it by decompling a class file. However, I have no time to disassemble the generated code today and tomorrow. [Here](https://stackoverflow.com/questions/25173208/value-integrity-guarantee-for-concurrent-long-writes-in-64-bit-openjdk-7-8) is an article that addresses the similar topic regarding `static long`. cc @rednaxelafx ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > Add benchmark for LongToUnsafeRowMap > ------------------------------------ > > Key: SPARK-26337 > URL: https://issues.apache.org/jira/browse/SPARK-26337 > Project: Spark > Issue Type: Test > Components: SQL > Affects Versions: 3.0.0 > Reporter: Liang-Chi Hsieh > Priority: Major > > Regarding the performance issue of SPARK-26155, I think it is better to add a > benchmark for LongToUnsafeRowMap which is the root cause of performance > regression. It can be easier to show performance difference between different > metric implementation in LongToUnsafeRowMap. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org