adriangb commented on PR #25338: URL: https://github.com/apache/datafusion/pull/25338#issuecomment-5734000496
Benchmark summary for `3af87370c1`. Every run compares it with the merge-base `64871d9`. [`projection_subquery`](https://github.com/apache/datafusion/pull/25338#issuecomment-5733944304), the suite this PR targets: | Query | `main` | this PR | | | --- | --- | --- | --- | | q01 bare `IN` | 244.3 ms | 1.15 ms | 213x | | q02 `COALESCE` over `IN` | 498.8 ms | 2.1 ms | 238x | | q03 correlated `IN`, equality | 4.2 ms | 1.62 ms | 2.6x | | q04 two `IN` columns | 368.8 ms | 1.88 ms | 196x | | q05 bare `NOT IN` | 240.2 ms | 1.12 ms | 214x | | q06 correlated `EXISTS` | 946 µs | 932 µs | 1.02x | | q07 correlated `IN`, non-equality | 100.3 ms | 101.9 ms | 1.02x | q06 and q07 are the controls. q07 keeps the three-join materialization, which this PR does not change, and https://github.com/apache/datafusion/issues/25336 covers it. The other suites show no change: | Suite | Queries | Faster | Slower | Total `main` | Total this PR | | --- | --- | --- | --- | --- | --- | | [tpch sf1](https://github.com/apache/datafusion/pull/25338#issuecomment-5733699483) | 22 | 0 | 0 | 770.0 ms | 769.4 ms | | [tpcds sf1](https://github.com/apache/datafusion/pull/25338#issuecomment-5733722109) | 99 | 1 | 3 | 9522.8 ms | 9340.5 ms | | [clickbench partitioned](https://github.com/apache/datafusion/pull/25338#issuecomment-5733757105) | 43 | 0 | 0 | 19328.0 ms | 19377.0 ms | The counts are from the min table. The four tpcds queries that move are q22 (1.05x slower), q43 (1.08x), q45 (1.05x) and q72 (1.11x faster), and the total is lower on the branch. The same comparison one run earlier gave 99 of 99 "no change", so these are the noise of this runner. -- 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]
