Yicong-Huang commented on PR #54555:
URL: https://github.com/apache/spark/pull/54555#issuecomment-4007364305

   
   > ```
   > batches = [batch1, batch2, batch3]
   > 
   > iterator = random choose one batch from batches
   > ```
   
   I am afraid this won't help as `peakmem` counts the overall memory used, by 
keeping a reference to all batches, they will be always counted in memory and 
we won't be able to see the real peakmem for iter-based UDF usage.
   
   That's the reason for replay from file when testing for `peakmem`. But then 
this could introduce disk I/O cost that will affect the `time*` measurement. 
   
   Let me try other ways, possibly try `mem*` instead of `peakmem*` to see if 
that helps.


-- 
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]

Reply via email to