andygrove opened a new pull request, #4919:
URL: https://github.com/apache/datafusion-comet/pull/4919
> This PR was created by an LLM as a draft PR. I will mark it as ready for
review after human review.
## Which issue does this PR close?
N/A — autonomous exploratory PR.
## Rationale for this change
Pad rows with one memcpy from a precomputed repeating pad buffer and write
pieces directly into the string builder instead of a scratch String, plus an
ASCII fast path and a fix for a rows*rows capacity hint in the column-length
path.
## What changes are included in this PR?
Pad rows with one memcpy from a precomputed repeating pad buffer and write
pieces directly into the string builder instead of a scratch String, plus an
ASCII fast path and a fix for a rows*rows capacity hint in the column-length
path.
## How are these changes tested?
Correctness: unit tests + seeded differential fuzz (bit-identical Arrow
output vs `main`).
Benchmark (criterion):
- spark_rpad_ default padding: 40.449% faster (base 130815ns -> cand 77902ns)
- spark_rpad_ length from column, multi-char padding: 41.531% faster (base
128111ns -> cand 74905ns)
- spark_lpad_ default padding: 37.778% faster (base 126685ns -> cand 78826ns)
- spark_lpad_ length from column: 31.923% faster (base 110963ns -> cand
75540ns)
- spark_lpad_ custom padding: 35.618% faster (base 123106ns -> cand 79258ns)
- spark_rpad_ multi-char padding: 52.516% faster (base 165109ns -> cand
78401ns)
- spark_lpad_ with truncation: 50.694% faster (base 110062ns -> cand 54268ns)
- spark_lpad_ multi-char padding: 57.246% faster (base 181356ns -> cand
77536ns)
- spark_rpad_ length from column: 32.23% faster (base 110303ns -> cand
74752ns)
- spark_rpad_ custom padding: 40.385% faster (base 130616ns -> cand 77866ns)
- spark_rpad_ with truncation: 50.556% faster (base 109976ns -> cand 54377ns)
Full criterion output:
```text
spark_lpad: default padding
time: [78.908 µs 79.083 µs 79.261 µs]
change: [−37.968% −37.778% −37.621%] (p = 0.00 <
0.05)
Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
6 (6.00%) high mild
spark_lpad: custom padding
time: [78.999 µs 79.093 µs 79.196 µs]
change: [−35.699% −35.618% −35.554%] (p = 0.00 <
0.05)
Performance has improved.
spark_rpad: default padding
time: [78.086 µs 78.410 µs 78.731 µs]
change: [−40.596% −40.449% −40.294%] (p = 0.00 <
0.05)
Performance has improved.
Found 20 outliers among 100 measurements (20.00%)
3 (3.00%) high mild
17 (17.00%) high severe
spark_rpad: custom padding
time: [77.806 µs 77.855 µs 77.931 µs]
change: [−40.474% −40.385% −40.290%] (p = 0.00 <
0.05)
Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
3 (3.00%) high mild
3 (3.00%) high severe
spark_lpad: multi-char padding
time: [77.424 µs 77.539 µs 77.677 µs]
change: [−57.312% −57.246% −57.187%] (p = 0.00 <
0.05)
Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
3 (3.00%) high mild
spark_rpad: multi-char padding
time: [78.306 µs 78.409 µs 78.530 µs]
change: [−52.583% −52.516% −52.442%] (p = 0.00 <
0.05)
Performance has improved.
Found 7 outliers among 100 measurements (7.00%)
6 (6.00%) high mild
1 (1.00%) high severe
spark_lpad: with truncation
time: [54.249 µs 54.317 µs 54.393 µs]
change: [−50.755% −50.694% −50.637%] (p = 0.00 <
0.05)
Performance has improved.
Found 13 outliers among 100 measurements (13.00%)
1 (1.00%) low mild
11 (11.00%) high mild
1 (1.00%) high severe
spark_rpad: with truncation
time: [54.146 µs 54.203 µs 54.268 µs]
change: [−50.698% −50.556% −50.370%] (p = 0.00 <
0.05)
Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
3 (3.00%) high severe
spark_lpad: length from column
time: [75.318 µs 75.499 µs 75.686 µs]
change: [−32.039% −31.923% −31.795%] (p = 0.00 <
0.05)
Performance has improved.
spark_rpad: length from column
time: [74.721 µs 74.759 µs 74.818 µs]
change: [−32.298% −32.230% −32.168%] (p = 0.00 <
0.05)
Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) high mild
2 (2.00%) high severe
spark_rpad: length from column, multi-char padding
time: [74.876 µs 74.937 µs 75.016 µs]
change: [−41.976% −41.531% −41.269%] (p = 0.00 <
0.05)
Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
3 (3.00%) high mild
2 (2.00%) 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.
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]
