Rangsh commented on PR #12172: URL: https://github.com/apache/seatunnel/pull/12172#issuecomment-5570595247
@nzw921rx Thanks for the review and for catching the Java 11 CV / Error trend. I agree that even though the absolute values are still low, a consistent increase across all four methods is worth checking more carefully before we treat it as noise. I will: 1. Add a few more ABBA runs (`baseline -> PR -> PR -> baseline`) for `DebeziumJsonFormatBenchmark` on Java 11 under the same JMH settings, and report Score / Error / CV / `gc.alloc.rate.norm` for each method. 2. If the higher CV / Error remains reproducible, compare baseline vs candidate runtime profiles on Java 11 (allocation / GC first, then CPU / wall if needed) to see whether the stability change is measurement variation or something introduced by this optimization. I will post the follow-up results on this PR once the additional runs are ready. -- 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]
