Rachelint commented on PR #12395:
URL: https://github.com/apache/datafusion/pull/12395#issuecomment-2356681769
Run benchmark introduced in #12513, about 10~20% improvement for the long
string(64 bytes).
```
INPUT LEN <= 12/ltrim_string_view [size=1024, strlen=12]
time: [35.007 µs 35.072 µs 35.137 µs]
change: [-1.6770% -1.4474% -1.2319%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN <= 12/ltrim_string [size=1024, strlen=12]
time: [35.804 µs 35.917 µs 36.043 µs]
change: [-1.5580% -0.7707% -0.0633%] (p = 0.07 >
0.05)
No change in performance detected.
Found 2 outliers among 10 measurements (20.00%)
1 (10.00%) high mild
1 (10.00%) high severe
INPUT LEN <= 12/ltrim_large_string [size=1024, strlen=12]
time: [35.830 µs 35.915 µs 36.017 µs]
change: [-2.2042% -1.2332% -0.3885%] (p = 0.01 <
0.05)
Change within noise threshold.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string_view [size=1024, strlen=64]
time: [35.466 µs 35.678 µs 35.902 µs]
change: [-14.201% -13.527% -12.873%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string [size=1024, strlen=64]
time: [41.778 µs 42.015 µs 42.295 µs]
change: [+0.6244% +1.3472% +2.1009%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_large_string [size=1024, strlen=64]
time: [40.613 µs 40.727 µs 40.856 µs]
change: [-1.7323% -1.3444% -0.9250%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string_view [size=1024, strlen=15]
time: [35.666 µs 35.728 µs 35.790 µs]
change: [-4.3312% -3.8432% -3.4491%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string [size=1024, strlen=15]
time: [37.365 µs 37.470 µs 37.569 µs]
change: [-0.4183% -0.0383% +0.3199%] (p = 0.85 >
0.05)
No change in performance detected.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_large_string [size=1024, strlen=15]
time: [37.139 µs 37.206 µs 37.270 µs]
change: [-1.4396% -1.1749% -0.9169%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN <= 12/ltrim_string_view [size=4096, strlen=12]
time: [138.14 µs 138.44 µs 138.71 µs]
change: [-2.4011% -1.6719% -1.0145%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN <= 12/ltrim_string [size=4096, strlen=12]
time: [140.09 µs 140.49 µs 140.88 µs]
change: [+0.5239% +0.9114% +1.2737%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN <= 12/ltrim_large_string [size=4096, strlen=12]
time: [140.43 µs 140.70 µs 140.95 µs]
change: [+0.4347% +0.9060% +1.2724%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string_view [size=4096, strlen=64]
time: [139.04 µs 139.37 µs 139.80 µs]
change: [-18.114% -17.283% -16.585%] (p = 0.00 <
0.05)
Performance has improved.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string [size=4096, strlen=64]
time: [164.84 µs 165.36 µs 166.00 µs]
change: [-1.8164% -1.2030% -0.5926%] (p = 0.00 <
0.05)
Change within noise threshold.
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_large_string [size=4096, strlen=64]
time: [163.68 µs 164.58 µs 165.49 µs]
change: [-3.6787% -2.6122% -1.5534%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string_view [size=4096, strlen=15]
time: [141.28 µs 141.55 µs 141.84 µs]
change: [-3.9775% -3.4188% -2.8950%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string [size=4096, strlen=15]
time: [146.82 µs 148.46 µs 151.08 µs]
change: [+0.3121% +1.4503% +3.2923%] (p = 0.07 >
0.05)
No change in performance detected.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high severe
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_large_string [size=4096, strlen=15]
time: [147.04 µs 147.51 µs 148.03 µs]
change: [-5.2701% -3.1076% -1.5059%] (p = 0.00 <
0.05)
Performance has improved.
INPUT LEN <= 12/ltrim_string_view [size=8192, strlen=12]
time: [276.36 µs 282.92 µs 295.77 µs]
change: [-2.9480% -0.3466% +4.3042%] (p = 0.90 >
0.05)
No change in performance detected.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high severe
INPUT LEN <= 12/ltrim_string [size=8192, strlen=12]
time: [280.94 µs 281.98 µs 283.42 µs]
change: [-0.1373% +0.6461% +1.4570%] (p = 0.16 >
0.05)
No change in performance detected.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high severe
INPUT LEN <= 12/ltrim_large_string [size=8192, strlen=12]
time: [281.94 µs 282.54 µs 283.15 µs]
change: [-10.644% -4.1065% -0.1338%] (p = 0.22 >
0.05)
No change in performance detected.
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string_view [size=8192, strlen=64]
time: [276.96 µs 277.37 µs 277.84 µs]
change: [-19.507% -19.151% -18.830%] (p = 0.00 <
0.05)
Performance has improved.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_string [size=8192, strlen=64]
time: [339.56 µs 342.24 µs 346.41 µs]
change: [-3.2469% -1.9321% -0.4932%] (p = 0.02 <
0.05)
Change within noise threshold.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high severe
INPUT LEN > 12, OUTPUT LEN > 12/ltrim_large_string [size=8192, strlen=64]
time: [337.07 µs 339.17 µs 342.10 µs]
change: [-2.4403% -1.5151% -0.4669%] (p = 0.01 <
0.05)
Change within noise threshold.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high severe
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string_view [size=8192, strlen=15]
time: [279.75 µs 280.22 µs 280.82 µs]
change: [-5.8159% -5.0886% -4.3076%] (p = 0.00 <
0.05)
Performance has improved.
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_string [size=8192, strlen=15]
time: [294.52 µs 295.12 µs 295.76 µs]
change: [-2.3012% -1.2254% -0.1972%] (p = 0.04 <
0.05)
Change within noise threshold.
INPUT LEN > 12, OUTPUT LEN <= 12/ltrim_large_string [size=8192, strlen=15]
time: [294.82 µs 296.32 µs 297.81 µs]
change: [-2.1467% -0.9477% +0.1515%] (p = 0.15 >
0.05)
No change in performance detected.
```
--
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]
