JakeDern commented on PR #10128:
URL: https://github.com/apache/arrow-rs/pull/10128#issuecomment-5406019272

   I actually think I found it - I stopped draining the buffers for the 
`Writer` implementation of `IpcRecordBatchSink` and I got a 13% bump in perf.
   
   I think dropping the Arcs one at a time as they're drained in the loop is 
much slower than clearing them all at the end for some reason? I don't really 
understand why because in my head it's the same number of atomic operations, 
but it seems to work.


-- 
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]

Reply via email to