Github user hhbyyh commented on the pull request:
https://github.com/apache/spark/pull/9878#issuecomment-159221643
I changed it to creating a new broadcast variable in each iteration.
I ran the benchmark with executor-core = 4 and iterations = 5. The result
is quite similar with the one before.
Memory usage decrease from 55% to 29%
time cost decrease from 524 to 304.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
