[GitHub] [arrow] ArgusLi commented on pull request #35574: GH-35559: [Java] Implementing JDBC Flight Stream Result Set asynchronous VectorSchemaRoot Producer

["ArgusLi (via GitHub)" <gi...@apache.org>]

ArgusLi commented on PR #35574:
URL: https://github.com/apache/arrow/pull/35574#issuecomment-1548752011

   > Is 0.5% performance improvement noticeable or even statistically significant?
   
   @lidavidm the posted results were taken from an automated test that we've discovered may not be the best indicator of performance. A more recent manual test fetching 500k rows has shown that this change led to a 4% faster performance - 114,256 ms vs 118,706 ms.
   
   The performance enhancement is iterative, we wanted to create this PR early to engage the community and get feedback as early as possible. More results are pending as we are working with a customer to see results in their lab. Our customer has a use case that leads to a fast consumer, they move millions of rows over the network back and forth, so there was a hypothesis that implementing double buffering would lead to a performance increase.


-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org