asolimando commented on PR #26094: URL: https://github.com/apache/datafusion/pull/26094#issuecomment-6063019356
Thanks @zhuqi-lucas for re-running the benchmark! Here is how I read the second run: - `physical_plan_tpcds_all`: 649.5 ms -> 636.3 ms (-2.0%) - `physical_plan_tpch_all`: 43.5 ms -> 42.6 ms (-2.1%), and most TPC-H queries are faster (up to -5%) - ClickBench (one table, no joins) does not change `optimizer_select_all_from_1000` is 5% slower, but it only does logical planning, which this PR does not change. On TPC-DS we save less time than in the first run (13 ms instead of 25 ms). I checked the commits merged to `main` in between and I don't see anything that would change how statistics are computed, and on TPC-H we save the same time as before (0.9 ms vs 1.0 ms). As `optimizer_select_all_from_1000` also moves by 5% without being related, I think this is just normal variation between runs (the TPC-DS error bar for the branch is within 5.3 ms drift in this run, which seems like a lot). WDYT? -- 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]
