mikemccand commented on PR #16245: URL: https://github.com/apache/lucene/pull/16245#issuecomment-5178256686
> > Maybe rerun with single segment (force merge to 1)? At least then the results are comparable to one another, though less realistic for near-real-time cases. > > +1, I think the difference in segments counts comes from the fact that we create baseline index with `useCMS=True` and `mergePolicy="TieredMergePolicy"` in `example.py`, but create `candidate_index ` without these args which will use `LogDocMergePolicy`, `Use ConcurrentMergeScheduler: false` defaultly. > > I will add the same args to `candidate_index ` to check. Thanks @vsop-479! This sounds like a bug in luceneutil's `example.py`... both indexing runs should use the same parameters, by default? Also, how come this change showed such a speedup in indexing? (once you corrected for this benchmarking bug) That seems too good to be true? -- 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]
