alamb commented on PR #11802: URL: https://github.com/apache/datafusion/pull/11802#issuecomment-2270931270
> @alamb It mainly improve the short queries in `clickbench_partitioned`, and In my expectations, other cases should have no changes, maybe can try the `clickbench_partitioned` case? That is my expectation too > 🤔 But it is strange that it get slower in other cases (especially the not short queries in `clickbench_1`), I guess the reasons chould be: > > * Some changes in the latest `main` but not in this `branch` make a difference? (assuming the `main` branch here is the latest), the slowers in `clickbench_1` look like this? My script tries to control for this by comparing against `git merge-base` -- FWIW the script I am using is here https://github.com/alamb/datafusion-benchmarking/blob/main/compare_branch.sh > * Maybe measurement noise exists? The slower in `tpch_mem_sf1` may be due to this? Yes, maybe > > I am running `clickbench_1` in my local to try reproducing it. Than thank you. I also hope to spend some time shortly looking into this (and it looks like you have done some additonal work too) -- 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]
