devmadhuu commented on PR #10549: URL: https://github.com/apache/ozone/pull/10549#issuecomment-5001863059
> Hi @devmadhuu, sure, I have updated the comment which mentioned that in-memory benchmark with file-based derby's benchmark. > > Here is the benchmark run for 50K results : > > Same scenario for both: 50,000 tracked unhealthy containers, one steady-state scan with ~10% churn (≈3.5K changed, ≈1.5K recovered, ≈45K unchanged). Only the write operation is timed (both approaches read existing rows first in production). > Before (master: delete-all + re-insert) After (this PR: diff sync) > Row operations delete 50,000 + insert 48,500 (~98.5K) update 3,500 + delete 1,500 (~5K) > Time (file-based) ~4.7 s ~3.9 s > > So at a realistic 50K chunk the PR is ~17% faster (4.7 s → 3.9 s). Now the question is , is it really worth to do this effort and PR just for 17% perf improvement. We usually vouch for min 40% perf improvement for any considerable code change. cc: @jojochuang @smengcl @rakeshadr @adoroszlai -- 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]
