alamb commented on PR #11247: URL: https://github.com/apache/datafusion/pull/11247#issuecomment-2211217568
Here are my benchmark results ``` -------------------- Benchmark tpch_mem.json -------------------- ┏━━━━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓ ┃ Query ┃ main_base ┃ add_short_circuit ┃ Change ┃ ┡━━━━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩ │ QQuery 1 │ 200.29ms │ 209.62ms │ no change │ │ QQuery 2 │ 35.62ms │ 34.17ms │ no change │ │ QQuery 3 │ 60.02ms │ 62.15ms │ no change │ │ QQuery 4 │ 63.80ms │ 63.97ms │ no change │ │ QQuery 5 │ 96.94ms │ 102.43ms │ 1.06x slower │ │ QQuery 6 │ 14.70ms │ 14.40ms │ no change │ │ QQuery 7 │ 199.35ms │ 207.52ms │ no change │ │ QQuery 8 │ 40.29ms │ 40.32ms │ no change │ │ QQuery 9 │ 109.85ms │ 106.79ms │ no change │ │ QQuery 10 │ 103.39ms │ 107.30ms │ no change │ │ QQuery 11 │ 72.28ms │ 74.07ms │ no change │ │ QQuery 12 │ 45.57ms │ 48.88ms │ 1.07x slower │ │ QQuery 13 │ 78.70ms │ 81.60ms │ no change │ │ QQuery 14 │ 18.92ms │ 19.50ms │ no change │ │ QQuery 15 │ 30.10ms │ 30.37ms │ no change │ │ QQuery 16 │ 44.68ms │ 41.71ms │ +1.07x faster │ │ QQuery 17 │ 159.93ms │ 162.91ms │ no change │ │ QQuery 18 │ 476.45ms │ 457.49ms │ no change │ │ QQuery 19 │ 47.57ms │ 45.01ms │ +1.06x faster │ │ QQuery 20 │ 104.00ms │ 92.20ms │ +1.13x faster │ │ QQuery 21 │ 290.77ms │ 284.74ms │ no change │ │ QQuery 22 │ 22.52ms │ 22.87ms │ no change │ └──────────────┴───────────┴───────────────────┴───────────────┘ ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓ ┃ Benchmark Summary ┃ ┃ ┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩ │ Total Time (main_base) │ 2315.72ms │ │ Total Time (add_short_circuit) │ 2310.01ms │ │ Average Time (main_base) │ 105.26ms │ │ Average Time (add_short_circuit) │ 105.00ms │ │ Queries Faster │ 3 │ │ Queries Slower │ 2 │ │ Queries with No Change │ 17 │ └──────────────────────────────────┴───────────┘ ``` So similarly I don't see much better or worse However, I think the end to end benchmarks might not be as good to evaluate this change as some micro benchmarks -- for example maybe we could extend https://github.com/apache/datafusion/blob/main/datafusion/core/benches/filter_query_sql.rs with the type of predicates this optimization would help 🤔 -- 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: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
