alamb commented on issue #5230:
URL:
https://github.com/apache/arrow-datafusion/issues/5230#issuecomment-1435052501
Here are my performance results (I did not dig into this yet). Basically it
shows no speedup or slow downs
Methodology:
```shell
git checkout sort-preserve-row-encoding
git cherry-pick 322e92bea6e28ada9f8d57d9429748fb58b2a2a5
cargo bench -p datafusion --bench sort -- --save-baseline
sort-preserve-row-encoding
```
Results
```
critcmp main sort-preserve-row-encoding
group main
sort-preserve-row-encoding
----- ----
--------------------------
sort f64 1.02
655.8±233.27µs ? ?/sec 1.00 640.9±24.78µs ? ?/sec
sort f64 preserve partitioning 1.00
5.1±0.09ms ? ?/sec 1.02 5.2±0.10ms ? ?/sec
sort i64 1.05
599.4±420.69µs ? ?/sec 1.00 571.7±11.11µs ? ?/sec
sort i64 preserve partitioning 1.00
4.5±0.08ms ? ?/sec 1.02 4.6±0.09ms ? ?/sec
sort mixed tuple 1.00
597.4±27.16µs ? ?/sec 2.30 1376.1±69.26µs ? ?/sec
sort mixed tuple preserve partitioning 1.00
4.9±0.13ms ? ?/sec 2.30 11.3±0.39ms ? ?/sec
sort mixed utf8 dictionary tuple 1.00
640.5±356.39µs ? ?/sec 1.39 892.1±43.52µs ? ?/sec
sort mixed utf8 dictionary tuple preserve partitioning 1.00
5.1±0.12ms ? ?/sec 1.42 7.3±0.15ms ? ?/sec
sort utf8 dictionary 1.00
200.7±4.89µs ? ?/sec 1.00 200.4±4.35µs ? ?/sec
sort utf8 dictionary preserve partitioning 1.00
1767.3±218.98µs ? ?/sec 1.00 1758.6±96.26µs ? ?/sec
sort utf8 dictionary tuple 1.00
683.6±1094.18µs ? ?/sec 1.24 846.7±107.74µs ? ?/sec
sort utf8 dictionary tuple preserve partitioning 1.00
4.8±0.24ms ? ?/sec 1.33 6.4±0.19ms ? ?/sec
sort utf8 high cardinality 1.00
2.1±0.04ms ? ?/sec 1.01 2.1±0.06ms ? ?/sec
sort utf8 high cardinality preserve partitioning 1.01
17.0±0.34ms ? ?/sec 1.00 16.9±0.39ms ? ?/sec
sort utf8 low cardinality 1.00
1258.3±84.71µs ? ?/sec 1.02 1280.3±30.47µs ? ?/sec
sort utf8 low cardinality preserve partitioning 1.00
1264.9±119.28µs ? ?/sec 1.01 1278.6±52.69µs ? ?/sec
sort utf8 tuple 1.00
1127.2±41.81µs ? ?/sec 1.53 1723.5±59.90µs ? ?/sec
sort utf8 tuple preserve partitioning 1.00
9.4±0.26ms ? ?/sec 1.53 14.4±0.44ms ? ?/sec
alamb@aal-dev:~/arrow-datafusion$
```
I plan to next run a prof run to see where time is being spent
--
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]
