luomh1998 commented on issue #8417: URL: https://github.com/apache/incubator-gluten/issues/8417#issuecomment-3766333078
> > > > [@guobj](https://github.com/guobj) We have implemented SMJ in Gluten, and TPC-DS Q72 showed a 1.14x (21s vs 24s) improvement in our test environments. You might have a try. > > > > > > > > > [@JkSelf](https://github.com/JkSelf) so we could set `spark.gluten.sql.columnar.forceShuffledHashJoin` to false by default? > > > > > > [@Yohahaha](https://github.com/Yohahaha) SMJ is currently marked as experimental in Gluten 1.3 due to its performance being 1.3 times slower than SHJ. We plan to further optimize the performance of SMJ next. > > Hey, [@JkSelf](https://github.com/JkSelf) any updates for the optimization for SMJ? Is it suitable to disable `spark.gluten.sql.columnar.forceShuffledHashJoin` by default to align the behavior with the spark and avoid the OOM issue which caused by SHJ. i have same question, spark.gluten.sql.columnar.forceShuffledHashJoin can set false by default ? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
