Github user WeichenXu123 commented on the issue:
https://github.com/apache/spark/pull/18798
@thunterdb
1) The dataframe deserialize from binary data will add overhead, (maybe
there is compaction or not, it depends on the datatype, cc @liancheng ) about
1x performance in my test.
2) My test try 20 times, the first 5 times used as warm up, the numbers in
the table is the mean of the 15 times results. You can trust the result I think.
3) When the vector size is small (1~10), the bottleneck is at many small
java object creation I think. So the performance result is different than the
larger vector.
---
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]
