jhorstmann commented on pull request #8113:
URL: https://github.com/apache/arrow/pull/8113#issuecomment-687585539
I saw the biggest effect when doing lexicographic sorting and added some
benchmarks for sorting. On my laptop the run to run variation was much to high,
but on an ec2 m5a.xlarge the improvement is up to 20%:
```
Running
/home/ubuntu/arrow/rust/target/release/deps/sort_limit_query_sql-934168e24deec398
Gnuplot not found, using plotters backend
sort_and_limit_by_int time: [123.66 us 123.75 us 123.85 us]
change: [-18.110% -17.989% -17.840%] (p = 0.00 <
0.05)
Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
1 (1.00%) low mild
2 (2.00%) high mild
1 (1.00%) high severe
sort_and_limit_by_float time: [115.25 us 115.37 us 115.50 us]
change: [-15.139% -14.976% -14.812%] (p = 0.00 <
0.05)
Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
4 (4.00%) high mild
1 (1.00%) high severe
sort_and_limit_lex_by_int
time: [134.29 us 134.39 us 134.49 us]
change: [-21.700% -21.591% -21.485%] (p = 0.00 <
0.05)
Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
1 (1.00%) low mild
1 (1.00%) high mild
2 (2.00%) high severe
sort_and_limit_lex_by_string
time: [133.06 us 133.15 us 133.24 us]
change: [-21.846% -21.728% -21.610%] (p = 0.00 <
0.05)
Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
1 (1.00%) low mild
3 (3.00%) high mild
1 (1.00%) high severe
```
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
