alamb commented on PR #9729:
URL: https://github.com/apache/arrow-rs/pull/9729#issuecomment-4259290247

   > I don't know if the small input sizes are too small to give 
stable/meaningful results? I would imagine most "real life" arrow arrays will 
have O(100) or O(1000) entries, not O(10)?
   
   Yeah I think most benchmarks we have use something like 8K rows as that 
mirrors what happens in most real world systems. Otherwise the per-array 
overhead dominates the actual work
   
   Also I think we have found that if the iteration time per benchmark is less 
than a few hundred microseconds, the results are very suceptabel to system 
noise. 


-- 
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