peter-toth commented on PR #10473:
URL: https://github.com/apache/datafusion/pull/10473#issuecomment-2182987502
This PR is more or less ready for review, tests are passing except for the
MSRV.
I focused only on the 3 performance improvements and deliberately kept the
code as close to the original as possible to ease review. I plan to open a
follow-up PR to tackle remaining TODOs and refactor CSE code a bit for better
readability.
Local benchmarks shows good perfroamance improvements, but @alamb please
confirm with your standard setup.
```
ptoth@ptoth-MBP16 arrow-datafusion % critcmp main better-cse-identifier
group better-cse-identifier
main
----- ---------------------
----
logical_aggregate_with_join 1.00 529.3±8.70µs ?
?/sec 1.00 528.4±6.17µs ? ?/sec
logical_plan_tpcds_all 1.00 71.3±2.74ms ?
?/sec 1.06 75.4±16.47ms ? ?/sec
logical_plan_tpch_all 1.00 7.3±0.19ms ?
?/sec 1.09 7.9±2.34ms ? ?/sec
logical_select_all_from_1000 1.02 15.9±2.00ms ?
?/sec 1.00 15.6±1.27ms ? ?/sec
logical_select_one_from_700 1.00 399.9±3.73µs ?
?/sec 1.02 408.3±4.29µs ? ?/sec
logical_trivial_join_high_numbered_columns 1.00 422.7±92.70µs ?
?/sec 1.00 423.7±88.41µs ? ?/sec
logical_trivial_join_low_numbered_columns 1.00 380.5±5.11µs ?
?/sec 1.01 385.3±4.22µs ? ?/sec
physical_plan_tpcds_all 1.00 536.8±11.44ms ?
?/sec 1.09 586.2±20.88ms ? ?/sec
physical_plan_tpch_all 1.01 38.5±12.88ms ?
?/sec 1.00 38.1±0.59ms ? ?/sec
physical_plan_tpch_q1 1.00 1108.3±12.50µs ?
?/sec 1.43 1585.0±12.90µs ? ?/sec
physical_plan_tpch_q10 1.00 1598.6±219.13µs ?
?/sec 1.15 1842.3±280.13µs ? ?/sec
physical_plan_tpch_q11 1.00 1339.7±14.58µs ?
?/sec 1.15 1540.2±247.46µs ? ?/sec
physical_plan_tpch_q12 1.00 1117.3±63.22µs ?
?/sec 1.12 1252.1±17.87µs ? ?/sec
physical_plan_tpch_q13 1.00 769.6±19.54µs ?
?/sec 1.12 864.0±12.96µs ? ?/sec
physical_plan_tpch_q14 1.00 930.6±216.63µs ?
?/sec 1.16 1081.1±125.78µs ? ?/sec
physical_plan_tpch_q16 1.00 1331.4±38.36µs ?
?/sec 1.15 1534.8±261.21µs ? ?/sec
physical_plan_tpch_q17 1.00 1209.8±29.88µs ?
?/sec 1.09 1322.0±17.01µs ? ?/sec
physical_plan_tpch_q18 1.00 1532.7±240.18µs ?
?/sec 1.04 1593.0±20.68µs ? ?/sec
physical_plan_tpch_q19 1.00 2.5±0.08ms ?
?/sec 1.06 2.7±0.04ms ? ?/sec
physical_plan_tpch_q2 1.00 3.2±0.70ms ?
?/sec 1.06 3.4±0.72ms ? ?/sec
physical_plan_tpch_q20 1.00 1642.3±16.61µs ?
?/sec 1.12 1845.0±302.12µs ? ?/sec
physical_plan_tpch_q21 1.00 2.6±0.65ms ?
?/sec 1.07 2.7±0.64ms ? ?/sec
physical_plan_tpch_q22 1.00 1196.6±101.79µs ?
?/sec 1.08 1292.8±15.01µs ? ?/sec
physical_plan_tpch_q3 1.00 1105.8±28.61µs ?
?/sec 1.23 1358.9±236.19µs ? ?/sec
physical_plan_tpch_q4 1.00 852.5±13.28µs ?
?/sec 1.12 957.4±161.05µs ? ?/sec
physical_plan_tpch_q5 1.00 1685.6±23.24µs ?
?/sec 1.10 1857.1±21.46µs ? ?/sec
physical_plan_tpch_q6 1.00 642.0±188.39µs ?
?/sec 1.01 647.4±12.04µs ? ?/sec
physical_plan_tpch_q7 1.00 2.2±0.45ms ?
?/sec 1.08 2.4±0.54ms ? ?/sec
physical_plan_tpch_q8 1.00 2.7±0.05ms ?
?/sec 1.08 2.9±0.06ms ? ?/sec
physical_plan_tpch_q9 1.00 1964.3±55.93µs ?
?/sec 1.08 2.1±0.04ms ? ?/sec
```
--
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
